Note: Descriptions are shown in the official language in which they were submitted.
BACKC~ROUND
The field of nuclear medicine has long been concerned with techTligues of
diagnosis wherein radiopharmaceuticals are introduced into a patient and the
resultant distribution or concentration thereof, as evidenced by gamma ray in-
tensities, is observed or tracked by an appropriate system of detection. An
important advantage of the diagnostic procedure is that it permits noninvasive
investigation of a variety of conditions of medical interest. Approaches to thisinvestigative technique have evolved from early pioneer procedures wherein a hand-
held radiation counter was utilized to map body contained areas of radioactivity to
10 more current systems for simultaneously imaging substantially an entire, in vivo,
gamma ray source distribution. In initially introduced practical systems, scanning
methods were provided for generating images, such techniques generally utilizing a
scintillation-type gamma ray detector equipped with a focusing collimator which
moved continuously in selected coordinate directions, as in a series of parallelsweeps, to scan regions of interest. A drawback to the scalming technique resides in
the necessarily longer exposure times required for the derivation of an image. For
instance, such time elements involved in image development generally are overly
lengthy to carry out dynamic studies of organ funetion.
By comparison to the rectilinear scanner described above, the later developed
20 "gamma camera" is a stationary arrangernent wherein an entire region of interest is
imaged nt once. As initially introduced the stationary camera systems generally
utilized a larger diameter sodium lodide, NaI (T11 crystal as a detector in
combination with a matrix of photomultiplier tubes. A multiple channel collimator
is interposed intermediate the source containing subject of investigation and thiæ
scintillation detector crystal. When a gamma ray emanating from the region of
investigative interest interacts with the crystal, a scintillation is produced at the
point of gamma ray absorption and appropriate ones of the photomultiplier tubes OI
the mab~ix respond to lhe thus generated light to develop output signals. The
original position of gamma ray emanation is determined by position responsive
30 networks associated with the outputs of the matrix. For additional information
concerning such camera, see:
I. Anger, H.O., "A New Instrument For Mapping Gamma Ray Emit-
ters", Biology and Medicine Quarterly Report UCRIr3653, 1957.
A eontinually sought goal in the performance of gamma cameras is that of
achieving a high resolution quality in any resultant image. ~urther, it is desirable to
achieve this resolution in combination with concomitant utilization of a highly
-1- .
o~
versatile r~dionuclide or rudiolabel, 99m-Technetium, having a gamma ray or photon
energy in the region of 1~10 l~eV. A broadened clinical utility for the cameras also
may be realized through the use and image identification of radiopharmaceuticalsexhibiting more than one photon ener~y level. WIth such an arrangement, two or aplurality of diagnostic aspects simultaneously may be availed the operator. For
example, in carrying out myocardial imaging, the above-identified 99m-Technetiummight be utilized in conjunc tion with lll-lndium, the latter contributing photon
energy in the regions of 173 and 247 KeV. Similarly, 81-Rubidium, exhibiting photon
energy in the range of 350 KeV might be utilized in conjunction with 81-Krypton, the
10 latter having gamma ray energy at about 120 KeV. The noted dual energy
characteristic of lll-lndium also might be utilized to achieve two aspects of
diagnostic data.
The resolution capabilities of gamma cameras incorporating scintillation
detector crystals, inter aliat is limited both by the light coupling intermediate the
detector and phototube matrix or array as well as by scatter phenomena of the
gamma radiation witnessed emanating from within the in vivo region of investiga-tion. Concerning the latter scattering phenomena, a degradation of resolution
occurs from scattered photons which are recorded in the image of interest. Such
photons may derive from Compton scattering into trajectories wherein they are
20 caused to pass through the camera collimator and interact photoelectrically with
the cyrstal detector at positions other than their point of in vivo derivation. Should
such photon energy loss to the Compton interaction be less than the energy
resolutioll of the system, it will effect an off-axis recordation in the image of the
system as a photopeak photon representing false spatial information or noise. Assuch scattered photons record photopeak events, the noise increase and consequent
resolution quality of the eamera climinishes. For the noted desirable 140 KeV
photons, the scintillation detector type camera energy resolution is approximately 15
KeV. With this resolution, photons which scatter through an angle from 0 to about
70 will be seen by the system as such photopeak events.
A continuing interest in improving the resolution qualities of gamma cameras
has led to somewhat extensive investigation into imaging systems incorporating
relatively large area semiconductor detectors. Such interest has been generated
principally in view of theoretical indications of an order of magnitude improvement
in statistically limited resolution to provide significant improvements in imagequality. In this regard, for example, reference may be made to the following
publieations:
~a~3~
n. R.N. Beck, L.T. Zimmer, D.B. Charleston, P.B. Hoffer, N.
Lembares, "The Theoretical Advantages of Eliminating Scatter in
Imaging Systems", Semi-ductor Detectors in Nuclear Medicine,
(P.B. Hoffer, R.N. Beck, and A. Gottschalk, editors ,~iety of
Nuclear Medicine, New York, 1971, pp. 92-113.
III. R.N. Beck, M.W. Schuh, T.I). Cohen, and N. Lembares, "Effects of
Scattered Radiation on Scintillation Detector Response'1, Medical
Radioisotope Scinti~raphy, IAEA, Vienna, 1969, Vol. 1, pp. 595-616.
IV. A.B. Brill, J.A. Patton, and R.J. Baglan, "An Experimental Com-
parison Scintillation and Semiconductor Detectors for Isotope
Imaging and Counting", IEEE Trans. Nuc. Sci., Yol. NS-l9, No. 3, pp
197-19~, 1972.
V. M.M. Dresser, G.F. Knoll, "Results of Scattering in Radioisotope
Imaging" IEEE Trans. Nuc. Sci., Vol. NS-20, No. 1, pp. 266-270,
1973.
Particular interest on the part of investigAtors has been paid to detectors
provided as hybridized diode structures formed basically of germanium. To derivediscrete regions for spatial resolution of impinging radiation, the opposed parallel
surfaces of the detector diodes may be grooved or similarly configured to de~ine20 transversely disposed rows and columns, thereby providing identi~iable discrete
regions of radiation response. Concerning such approaches to treating the de-
tectorsS mention may be made of the following publications:
VI. J. Detko, "Semiconductor Dioxide M&trix for Isotope Localization",
Phys. Med. B_1., Vol. 14, No. 2, pp. 245-253,1969.
VII. J.F. Detko, "A Prototype, Ultra Pure Germanium Orthogonal Strip
Gamma Camera," Proceedings of the IAEA Symposium on Radi-
oisotope Scintigraphy, IAEA/SM-164/135, Monte Carlo, October
1972.
YlII. R.P. Parker, E.M. Gunnerson. J.L. Wankling, and R. Ellis, "A
Semiconductor Gamma Camera with Quantitative Output," Medical
adioisotope Scintigraphy.
I~. V.R. McCready, R.P. Parker, E.M. Gunnerson, R. Ellis, E. Mo~ss,
W.G. Gore, and J. Bell~ "Clinical Tests on a Prototype Semi-
conductor Gamma~amera," British Journal of Radiolo~, Vol. 44,
58-62, 1971.
X. Parker, R.P., E.M. Gunnerson, J.S. Wankling, R. Ellis, "A Semi-
conductor ~amma Camera with Quantitative Output," Medical
Radioisotope Scinti~raphp, Vol. 1, Vienna, IAEA, 1969, p. 71
Xl. Detko9 J.P., "R- Prototype, yltra-Pure Germanium, orthogonal-
Strip Gamma Camera," Medical Radioisotope Scintigra~z, VoL 1,
Vienna, IAEA, 1973~ P. 241.
Xll. Schlosser, P.A., D.W. Miller, M.S. Gerber, R.F. Redmond, J.W.
Harpster, W.J. Collis, W.W. Hunter, Jr., "A Practical Gamma Ray
Camera System Using High Purity Germanium," presented at the
1973 IEEE N~lclear Science Symposium, San Francisco, November
1973; also published in ~L~ I r-~ N ~ 1 5~i, Vol. NS-21, No. 1
Febrllary 1974, p. 658.
XIII. Owen, R.B., M.L. Awcock, "One and Two Dimensional Position
Sensing Semiconductor Detectors," IEEE Trans Nucl, Sci., Vol. NS-
51, June 1968, p. 290.
lQ In the more recent past, investigators have shown particular interest in forming
orthogonal strip matrix detectors from p-i-n semiconductors fashioned from an ultra
pure germanium material. In this regard reference is made to U.S. Patent No.
3,761,711 as well as to the following publications:
XIV. J.F. Detko, "A Prototype, Ultra Pure Germanium, Orthogonal Strip
Gamma Camera," ProceedinFs OI the IAEA SYmPosium on Radi-
oisotoDe Scinti~raphy, IAEA/SM-164/135, Monte Carlo, October,
1972.
XV. Schlosser, P.A., D.W. Miller, M.S. Gerber, R.F. Redmond, J.W.
Harpster, W.J. Collins, W.W. Hunter, Jr., "A Practical Gamma Ray
Camera System Using High Purity Germanium," presented at the
1973 IEEE Nuclear Science Symposium, San Francisco, November
1973; also published in IEEE Trans. Nucl. Sci., Vol. NS-21, No. 1,
February 1974, p. 658.
High purity germanium detectors promise numerous advantages both in gamma
camera resolution as well as practicality. For instance, by utilizing high purity
germanium as a detector, lithium drifting arrangements and the like for ~educingimpurity concentrations are avoided and the detector need only be cooled to
reguisite low temperatures during its clinical operation. Read-out from the ortho-
30 gonal strip germanium detectors is described as being carried out utilizing a number
o~ techniques, ior instance, each strip of the detector may be connected to the
preamplifier-~mplifier channel and thence directed to an appropriate logic function
and visual readout. In another arrangem ent, a delay line readout system is
suggested with the intent of r educing the number of preamplifiers-amplifier
channels, and a technique of particular interest utilizes a charge splitting method.
With this method or technique, position sensitivity is obtained by eonnecting each
contact strip of the detector to a charge dividing resistor network. Each end ofeach network is connected to a virtual earth, charge sensitive preamplifier. When a
gamma ray interacts with the detector, the charge released enters the string of
4û resistors and divides in relation to the amount of resistance between its entry point
in the string and the preamplifiers. Vtilizing fewer preamplifiers, the cost andcomplexity of such systems is ~dvantageously reduced. A more deta;led description
of this readout arrangement is provided in:
XVI. Gerber, M.S., Miller9 D.W., Gillespie, B., and Chemistruck, R.S., "Instrumetation For a High Purity Germanium Position Sensing
Gamma Ray Detector," IEEE Trans. on Nucl. Sci., Vol. NS-22 No. 1,
February, 1975, p. 416
To achieve requisite performance and camera image resolution, it is necessary
10 that substantially aU sources of noise or false information within the system be
accounted for. In the absence of adequate noise resolution, the performance of the
imaging systems may be compromised to the point of impracticality. Until the more
recent past, charge splitting germanium detector arrangements have not been
considered to be useful in gamma camera applications in consequence of thermal
noise anticipated in the above-noted resistor divider networks, see publication VIIg
supra. However, as will be evidenced in the description to follow, such considera-
tions now are moot.
Another aspect in the optimization of resolution of the images of gamma
carneras resides in the necessarily inverse relationship between resolution and
20 sensitivity. A variety of investigations have been conducted concerning this aspect
of camera design, it being opined that photon noise limitations, i.e. statistical
fluctuations in the image, set a lower limit to spatial resolution. Further, it has
been pointed out that the decrease in sensitivity witnessed in conventional highresolution collimators may cancel out any improvements sought to be gained in
image resolution. A more detailed discourse concerning these aspects of design are
provided, for instance, in the following publications:
XVII. E.L. Kel~er and J.W. (~oltman, "Modulation Transfer and Scintilla- -
tion Limitations in ~amma Ray Imaging" J. Nucl. Med. 9, lû, 537-
3~ 545 11968)
XVIII. B. Westerman, R.R. Sharma, and J.~. Fowler, "Relative Im-
portance of Resolution and Sensitivity in Tumor Detection", a.
Nucl Med. 9,12 638-640 (1968)
Generally, the treatment of the signals derived at the entrance detec$ion
portion of gamma cameras involves a form of spatial or coordinate identification of
photons reaching the detector and additionaLly, a form of analysis of the energy of
radiation reaching the detector. Spatial analysis may be carried out by difference
summing circuits, while energy determination is carried out by additive summing
circuits. Further pulse height analy~ers may be utilized as one discriminating
component of a system for determining the presence of true or false imaging
information. In any of the systems both treating noise phenomena and seeking a
high integrity of spatisl information, a control is required which carries out
appropriate noise filtering while segregating true from false information. In
addition to the fore~oing, it is necessary that the 1'throughput rate" of the system be
maximized in order that it may accommodate a highest number of bits or pulses
representing spatial and energy data.
Another operational phenomenon tending to derogate from the spatial resolu-
10 tion quality performance of the cameras is referred to as "aliasing". This
phenomenon represents a natural outgrowth of the geometry of the earlier-noted
orthogonal strip germanium detector. A more detiled discussion of this aspect ofthe gamma cameras is provided at:
XIX. J.W. Steidley, et al., "The Spatial Frequency Response of Ortho-
gonal Strip Detectors: IEEE Trans. Nucl. Sci., February, 1976.
To remain practical, it is necessary that the imaging geometry of stationary
type gamma cameras provide for as large a field of v;ew as practical. More
particularly, such considerations require a camera field of view large enough to~0 encompass the entire or a significant extent of the profiles OI var;ous organs of
interest. Because of limitations encountered in the manufacture of detector
crystals, for instance, high purity germanium crystals, the size of solid state
detector ~omponents necessarily is limitedO As a conseguence, composite detectorconfigurations are required which conjoin a plurality of smaller detector components
to provide an imaging field of view or radiation acceptance geometry of effectively
larger size. However, such union of a multitude of detector components must be
carried out without the concurrent generation of noise phenomena and without a
significant loss of image information validity and acuity. For instance, in the latter
regard, spatial information must have a eonsistency of meaning across the entire3û extent of an ultimately displayed image of an organ, otherwise, clinical evaluation
of such images may be encumbered. Preferred arrangements for intercoupling the
discrete detector components within an overall array thereof is described in United
States Patent by M.S. Gerber and D.W. Miller, entitled "Gamma Camera System",
Patent No. 4,061,919.
~3~
The control systems utilized with garnma cameras having multi-component
detectors further are called upon to collect image data therefrom at an optimum
rate while evaluating the validity thereof and assigning it an appropriate address
function. Such address assignment may vary in nature depending upon the selectedmode of circuit interrelationship of the discrete detector components with the
array. An addition~l function of the control system is to identify the spatial
position of the detector-photon interaction for select but different energy levels.
This requires a technique for normalizing the spatial labels of such signals while
properly evaluating the energy level states thereof as representing valid image
10 information. The rapidity with which this data is treated, as by assigning spatial
regional factors to it, as well as evaluating it for validity becomes a particularly
important aspect of the control systems where they are contemplated for use in
clinical dynamic function studies. With such studies, dynamic alterations in an
image component occurring within any segment of the image area should be
followed closely in correspondence with the actual movement of the image source.Accordingly, efficient image signal treatment by the camera system is required.
-6A-
SV MM~RY
The invention provides, in a system for imaging the distribution within a regionof interest of isotopic materials emitting radiation of given photon energy, said
system including composite solid state detector means having a plurality of discrete
componer~ts which are operationally associated to provide spatial coordinate para~
meter outputs representative of the spatial disposition of corresponding interactions
of said radiation with said detector means, the improvement comprising:
first output treating means connected to receive said spatial coordinate
parameter outputs of said detector means components, actuable to selectively ~ilter
10 and sum said outputs to derive corresponding coordinate channel signals and an
energy channel signal having values corresponding respectively with said spatialdisposition and given photon energy exhibited at a said interaction, said first output
treating means further including control means for effecting said actuation to filter
and sum and for deriving a ~ata acceptance signal in correspondence with said
coordinate and energy channel signals, said control means being responsive to a
received reset signal to reset said first output means to a clear condition;
means including spatial coordinate multiplexer means and energy channel
multiplexer means respectively coupled to be addressed by sa;d energy channel
signals, each said multiplexer means being responsive to a coded actuating signal to0 effect a transference of said channel signals addr~ssed thereto;
process control means including memory means for receiving said data
acceptance signals and selectively retaining them in received serialized fashion, and
actuable to derive said coded actuating signal in correspondence with said serialized
data acceptance signals;
seguential control means for selectively actuating said process control
means and regulating an operational cycle of said system; and
second treating means responsive to said transfered channel signals for
deriving readout information representative thereof.
The invention also provides, in a system for imaging the distribution within a
30 region of interest of isotopic materials emitting radiation exhibiting two or more
levels of photon energy, said system including composite solid state detector means
having a plurality of discrete components which are operationally associated to
provide spatial coordinate parameter outputs representative of the spatial disposi-
tion of corresp~nding interactions of said radiation with said detector means, the
improvement csmprising: ~
~3~
~ i~rst outE~ut treating menns connected to receive said spatial coordinate
parameter outputs of snid detector means components, actuable to selectively filter
and sum said outputs to derive corresponding coordinate channel signals and an
energy channel signal having values corresponding respectively with said spatialdisposition and the level of said photon energy exhibited at a said interaction, and
including control means for effecting said actuation to filter and sum and for
deriving a data acceptance signal in correspondence with said coordinate and energy
channel signals, said control means being responsive to reset said first output means
to a clear condition;
means including spatial coordinate multiplexer means and energy channel
multiplexer means respectively coupled to be addressed by said coordinate channel
and said energy channel signals, each said multiplexer means being responsive to a
coded actuating signal to effect a transferen¢e of said channel signals addressed
thereto;
process control means including means for receiving said data accept-
ance signals to effect a de-randomization thereof and actuable to provide a saidcoded actuating signal in correspondence with a said data acceptance signal;
second treating means including divider network means responsive to a
said energy ch~nnel multiplexer means transfersed energy channel signal and to said
20 spatial coordinate multiplexer means transferred signals for normalizing said spatial
coordinate channel signals with respect to the said photon energy related value of
their corresponding said energy channel signal, said second treating means beingconfigured for deriving readout information corresponding with said normalized
spatial coordinate channel signals; and
sequential control means for selectively actuating said process control
means and regulating an operational cycle of said system.
The invention also provides, in a system ~or imaging the distribution within a
region of interest of isotopic materials emitting radiation of given photon energy or
energies, said system including composite solid state detector means having à
30 plurality of discrete components whch are operationally associated within pre-
determined portions of said detector means to provide spatial coordinate parameter,
x-, y-designated outputs representative of the spatial disposition of corresponding
interations of said resdiation with said detector means, the improvement compris-
ing:
first output treating means conneeted to receiYe said sp~tial coordinate
parameter, X-9 y-designated outputs of said cornponents of each said predetermined
~a~3~
portion of said detector means, actuable to selectively filter and sum said outputs
when operating from a clear condition to derive corresponding x- and y-designated
coordinate channel signals and an energy channel signal having values corresponding
respectively \Nith said spatial disposition and given energy exhibited at a saidinteraction, said first output treating means including first evaluating means
responsive to the value of said given energy equaling or exceeding a predetermined
value to derive a select output, and further including control means responsive to
the presence of said evaluating means select output for effecting said actuation to
filter and sum and for deriving a data acceptance signal in time corresponding with
10 said control means being responsive to a reset signal when submitted thereto to
reset said first output treating means to said clear condition;
means including x-position multiplexer means, y-position multiplexer
means and energy multiplexer means respectively coupled to be addressed by said x-
and y-designated coordinate and energy channel signals, each said multiplexer
means being responsive ~o a coded actuating signal to transfer said channel signals
addressed therto;
storage means, having receive and hold modes, for receiving each
transIerred said channel signal then in said receive mode, and actuable to assume
said hold mode re taining each said channel signal over a given interval, said storage
20 means having outputs for asserting each said retained channel signal;
process control means including asynchronous memory means for accept-
ing and retaining said data acceptance signals in received serialized fashion and
actuable to provide said coded actuating signal;
sequential control means for selectively actuating said process control
means to effect selective transfer of said channel signals to said storage means, and
for actuating said storage means to derive and retain said hold mode for said given
interval, and
second treating means responsive to said channel signals asserted at said
storage means outputs for deriving readout information representative thereof.
Finally, the invention provides, in a system for imaging the distribution withina region of interest of isotopic material emitting radiation of given photon energy,
said system including composite solid state detector means having a plurality ofmutually adjacently disposed discrete detector components each having one of twooppositely disposed charge csllecting surfaces positioned substantially within acommon plane for e~posure to said radiation, said components being arranged to
estab]ish linearly oriented groupings of respective said opposed surfaces, each said
~L3~
groupings of surfaces being electric~lly intercoupled and associated with impedance
means for providing x- and y-designated coordinate outputs from respective
mutually orthogonally aligned and oppositely disposed ones of said groupings
a~ssociated with a common said detector component at which an interaction with
said radiation corresponding with said coordinate output occurs, the improvementcomprising:
x-coordinate output treating means responsive to said x-designated
coordinate OUtplltS, actuable to selectively filter and sum said outputs from a clear
condition to derive corresponding x-designated coordinate channel signals and an10 energy signal having values corresponding respectively with a coordinate aspect of
the spatial location of said interaction and the value of said photon energy thereof,
and including control means for effecting said actuation to filter and sum and for
deriving an x-data signal corresponding with said channel and energy signals, said
control means being responsive to a reset signal when submitted thereto to resetsaid treating means to said clear condition;
y-coordinate output treating means responsive to said y-designated
eoordinate outputs, actuable to seleetively ilter and sum said outputs from a elear
eondition to derive corresponding y-designated eoordinate ehannel signals and anenergy signal hav;ng values eorresponding respeetively with a eoordinate aspeet of
2û the spatial location of said interaetion and the value of said photon energy thereof,
and ineluding eontrol means for effeeting said aetuation to filter and sum and for
deriving a y-data signal eorresponding with said ehannel and energy signals, said
eonkol means being responsive to a reset signal when submitted thereto to reset
said treating means to said elear eondition;
means ineluding x-position multiplexer means, y-position multiplexer
means and energy multiplexer means, respeetively eoupled to be addressed by saidx- and y-designated eoordinate ehannel signals and at least one said energy signal,
eaeh said multiplexer means being responsive to a eoded aetuating signal to transfer
said signals addressed thereto;
proeess eontrol means ineluding eoineidence network means responsive
the said x-data and y-data signals eorresponding with a given said interaetion for
deriving a pair eode signal, and axynehronous memory means responsive to said pair
eode signal for aeeepting and retaining said x- and y-data sign~ls in reeeived
serialized fashion and aetuable to provide and assert said eoded aetuating signal at
said multiplexer means;
--10--
~L3~
menns responsive to said x- and y-designRted coordinate channel signals
when transferred from said x--position multiplexer means, said y-position multi-plexer rmeans, and said energy multiplexer means for deriving readout information
representative thereof; and
sequential control means for selectively actuating said process control
means to effect said transfer of said signals.
For a fuller understanding of the nature and the objeet of the invention,
reference should be had to the following detailed description taken in conjunction
with the accompanying drawings.
Bl~IEF DESCRIPTION OF THE DRAWINGS
Pigure 1 is a schematic representation of a gamma camera arrangement as
may utilize the improvement of the invention, showing, in block schematic form,
general control functions;
Figure 2 is a pictorial representation of a solid state orthogonal strip high
purity germanium detector component incorporating a charge splitting resistor
network in combination with preamp~ification electronics;
Figure 3 is a schematic representation of a solid stage strip detector and a
schematic collimator functionally associated therewith as such system componentsrelate to the radiation source within a region of clinica1 interest;
Figures 4(a)-4(c) are a schematic and graphical representation of the funda-
mental geometry associated with the interrelationship of a multi-channel collimator
and a solid state detector;
Figure 5 is a pictorial representation of a collimator array which may be
utili~ed with the system of the invention;
Figure 6 is a pictorial view.of two internested members of the collimator of
Figure 5;
Figures 7(a)-7(c) respectively and schematically depict representations of a
source distribution as related with the gqometry of an orthogonal strip detector and
image readouts for illustrating aliasing phenomena;
Figures 8~a)-8(d) portray ~ertically aligned graphs relating modulation transferfunction with respect to resolution as such data relates to aliasing phenomena,
Figure 8(a) showing collimator modulation transfer function (MTFC) with FWHM
resolution of 1.33.~ Figure 8(b) showing a consequent alias frequency spectrum which
is processed by the electronics of the camera system, Figure 8(c) showing electronic
MTF for given resolutions, and Figure 8(d) showing camera system MTF's revealingaliasing introduced by the orthogonal strip solid state detector;
~3~
Figures 9(a)-9(d) provide curves showing the results of aliasing correction as
compared with the curves of figures 8(a)-8(d), Figure 9(a) looking to collimatordesign as an antialiasillg filter, Figure 9(b) showing a consequent aliasing frequency
spectrum which is processed by the electronics of the system, Figure 9(c) showing
the consequence of electronics used for antialiasing post-filtering, and Figure 9(d)
showing total system MTF revealing the elimination of aliasing phenomena;
Figure 10 is an equivalent noise model circuit for solid stage detectors as
utiliæed in accordance with the instant invention;
Figure 11 is a circuit model of a detector compnent and related resisto
10 network, schematically representing a position-sensitive detector arrangement;
Figure 12 is a block schematic diagram of a gamma camera control system
configured as it is related to a single detector component output;
Figure 13 is a schematic block diagram of a gated integrator configuration
which may be utilized with the instant invention;
Figure 14 is a schematic circuit representation of the configuration described
in connection with Figure 13;
Figure 15 is a schematic representation of the logic components of a control
arrangement which may be utili~ed with the systern of the invention;
Figure 16 is a circuit timing diagram corresponding with the schematic
20 representation shown in Figure 15;
Figure 17 is a pictorial and schematic representation of an array of detector
components showing the interconnections thereof to form a composite detector or
region thereof as may be utilized with the systerm of the invention;
Figure 18 is a block schematic representation of the control system utilized to
receive and treat the outputs of the detector array configuration of Figure 17;
Figure 19 is a block schematic diagram of an embodiment of the control
system of the invention as it is utilized for treating the signals developed by the
control arrangement of Figure 18;
Figure 20 is a schematic and pictorial representation of another array of
30 detector components, interconnected in accordance with a "row-column" readout geometry;
Figure ~1 is a schematic and pictorial representation of another array of
detector components, each of which is formed associated with a surface type
impedance arrangement, the components being interconnected in the noted "row-
column" fashion;
--12--
6~
Figure 22 is a schematic and pictorial representntion of another array of
detector cornponents interconnected in accordsnce with the noted "row-column"
geometry;
Figure 23 is a block schematic di~gram of a control system utilized in treating
one spatial channel output of the noted "row-column" detector component inter-
connection geome*y;
Figure 24 is ~ schematic block diagram of a control circuit operating in
conjunction and cooperation with the control system of Figure 23; and
Figure 25 is a block diagram of a control arrangement for utilization with the
10 noted "row-column" interconnection of detector components, the figure representing
an alternate control arrangement within the diagram of Figure 19.
DETAILED DESCRIPTION
In the discourse to follow, the control system of the invention initially is
described in conjunction with the arrangements utilized for physically acceptinggamma radiation from a clinically determined region of interest. In particular,
initial acceptance teehniques for collimating such radiation as well as pararneters
required for such collimation are set forth. Following that discussion, the discourse
sets forth techniques for achieving optimized system performance with respect tonoise characteristics which otherwise would be encountered with the solid state
20 detector arrangement of the invention. Looking additionally to techniques forimproving through-put rate characteristics for tlle system, the discussion initially is
concerned with a control over a detector arrangement incorporating only a one
detector component. Following this basic description, however, preferred teehniques
are set forth for associating a plurality of solid state detector components within Q
predetermined array or mosaic configuration. Such con-figurations and operatoinal
criteria therefore being established, the discussion then looks to a control system
which may operate with radiopharmaceutical sources of more than one detectible
energy level and which serves to treat resulting signals as well as label and address
them to achieve practical overall imaging fields of view which maintain efficient
30 signal treatment.
As indicated in the foregoing, during contemplated clinical utilization, a
gamma camera arrangement according to the instant invention is used to image
gamma radiation within patients. Looking to Fig. 1, an exaggerated schematic
representation of such a clinical environment is revealed generally at 10. The
environment schematically depicts the cranial region 12 of a patient to whom has
--13--
been administered a radi~labeled pharm~ceutic~l, which phar~naceutical will havetended to concentrate within a region of investigative interest. Accordingly,
radiation is depicted as emanating from region 12 as the patient is positioned on
some supporting pl~tform 14. Over the region 12 is positioned the head or housing 16
of a gamma camera. Extending outwardly from the sides of housing 18 are mountingflanges, as at 18 and 20, which, in turn, may be connected in pivotal fashion with an
appropriate supportingr assembly (not shown). Housing 16 also supports a vacuum
chamber 22 defined by upper and lower vacuum chamber plates shown) respectively,at 24 and 26 conjoined with an angularly shaped side defining flange member 28.
10 Lower vacuum chamber plate 269 preferably, is formed of aluminum and is
configured having a thin entrance window portion 30, directly above which is
provided an array of discrete solid state detector components, as shown generally at
32. Array 32, in turn, ;s operationally associated with the "cold finger" component
34 of an environmental control system, which preferably includes a cryogenic region
refrigerating unit of a closed-cycle variety, shown generally at 36. An ion pump, as
at 38, assures the integrity of the vacuum in chamber 22, such pump, in conjunction
with the refrigerating unit 36, being mounted for association with chamber 22
through upper vacuum plate 24, the latter which may be formed, for instance, of
stainless steel. Vacuum pum~down of the chamber 22 is accomplished by first using
20 a sorption-type roughing pump, then using the ion pump shown to reduce and
maintain the ~hamber pressure at 10 6 Torr sr less.
Electronics incorporated within chamber 22 include preliminary stages of
amplification, for instance field effect transistors (FErs) as at 40 which are
mounted upon a plate 42 coupled, in turn, betwleen cold-finger 34 and side channel
28. Thus connected, the plate 40 evidences a temperature gradient during the
operation of the unit which provides a selected ideal temperature environment ofoperation for the amplification stages. The outputs of these stages are directedthrough subseguent stage electronics, shown within a housing 44, which, in turn,provides elec~ ical communication to externally disposed control electronics through
30 conduit 46 and line 48. To provide for appropriate operation, chamber 22 generally
is retained at a temperature of, for instance, about 77K, while the FET's, 40,
mounted upon plage 42, are retained at about 130K to achieve low noise perform-ance.
Mounted outwardly of window portion 30 an in alignment with the detector
array 32 is a collimator, shown generally at 50. During the operation of the gamma
camera, radiation emanating from source 12 is spatiaUy coded initially a$ collimator
--14-
~l~3~
5û by attenuating or rejecting off-axis radiation representing false image informa-
tion. ThRt radiation passing ~ollimator 50 impinges upon detector array 32 and asignificant portion thereof is converted to discrete charges or image signals.
Detector array 32 is so configured as to distribute these signals to resistor chains as
well as the noted pl~eamplification stages 40 retained within chamber 22 to provide
initial signals representative of image spatial information along conventional
coordinate axes as well as representing values for radiation energy levels. This data
then is introduced, as represented schematically by line 48, to filtering and logic
circuitry which operates thereupon to derive an image of optimized resolution and
l~û veracity. In the latter regard, for instance, it is desired that only true image
information be elicited from the organ being imaged. Ideally, such information
should approach the theoretical imaging accuracy of the camera system as derived,
for instance, from the geometry of the detector structure 32 and collimator
arrangement 33 as well as the limitations of the electronic filtering and control of
the system.
Image spatial and energy level signals from line 48 initially, are introduced
into Anti-Symmetric Summation and Energy Level Derivation functions represented
at block 52. As is described in more detail later herein, the summation carried out
at block 52 operates upon the charges directed into the resistive chains or networks
20 associated with the orthogonal logic structuring of detector array 32 to derive
discrete signals or charge values corresponding with imsge element location.
Additionally, circuitry of the function of bloclc 52 derives a corresponding signal
representing the energy levels of the spatial information. The output of block 52 is
directed to Filtering Amplification and Energy Discrimination functions as are
represented at block 54. Controlled from a Logic Control function shown at block56, function 54 operates upon the signal input thereto to accommodate the systemto parallel and series defined noise components through the use of Gaussian
arnp]ification or shsping, including ~apezoidal pulse shaping of data representing
the spatial location of image bits or signals. Similarly, the energy levels of
30 incoming signals are evaluated, for instance, utilizing for instance multlple channel
analyzer components controlled by logic circuitry at 56 to establish energy level
windows for data received within the system. In this regard, signals falling above
and below predetermined energy levels are considered false and are blocked. ~romAmplification and Discrimination stage 54 and I,ogic Control 56, the analy~ed
signals are directed into an Information Display and Readout Function, as repre-sented at block 58. Components within function block 58 will include display
--15--
~L3~
screens of various configurat;ons, ;magc recording devices, for instance, photo-grflphic app~ratus of the h~stAnt developing variety, radiation readout devices and
the like, which are controlled at the option of the system operator.
As outlined above, the instant description now looks in more detail to the
configuration of the collimator structure 50. To facilitate such description,
however, the structure of a single component within the detector array 32 is
described in conjunction with Fig. 2. 1ater dis~ussion and figures will reveal the
interrelationships of such impedance networks and their equivalents as they are
operatively associated witn a multi-component detector array. Looking to that
10 figure, an e~caggerated pictorial representation of such a component of the detector
array is revealed at 60. Detector component 60 may be fabricated from p-type high
purity germanium by depositing an n-type contact on one face and a p-type contact
on the opposite face of a rectangular planar crystal. Accordingly, a high puritygermanium region of the crystal, as at 62, serves as an intrinsic region betweentype semiconductor region contacts 64 and n-type semiconductor region contaets as
at 66. The intrinsic region 62 of the ~i-n detector components forms a region
which is depleted of electrons and holes when a reverse bias is applied to the
contacts. Grooves as at 68a-68c are cut into the continuous p-type contact or
region at one face of the component to form strips of isolated p-type semiconductor
20 material. On the opposite face of the detector component, orthogonally disposed n-
type semiconductor strips similarly are formed through the provision of grooves 70a-
70c. Configured having this geometry, the detector component 60 generally is
referred to as an orthogonal strip detector or an orthogonal strip array semi-
conductor detector component. The electrode strips about each of the opposed
surfaces of compnent 60, respectively, are connected to external charge splitting
resistor networks revealed generally at 72 and 74. Resistor network 72 is formed of
serially coupled resistors 76a-76e which, respectively, are tapped at their regions of
mutual interconnection by leads identified, respectively, at 78a-78d extending, in
turn, to the orthogonfll strips. The opposed ends of network 72 terminate in
30 peramplification sta~es %0 and 82, the respective outputs of which, at 84 and 86,
provide spatial output data for insertion within the above-described summation and
energy level derivation function 52 to provide one detector component orthogonal or
coordinate output, for instance, designated as a y-axis signal.
In similar fashion, network 74 is comprised of a string of serially coupled
resistors 88a-88e, the mutual interconnections of which are coupled with the
electrode strips at surface 66, respectively, by leads 9ûa-gOe. Additionally,
--16-
prenmplification stnges as at 92 and 94 provide OlltpUtS, respect;vely, at lines 96 and
98 carrying spatial data or signals representative of image information along an x
axis or axis orthogorlally disposed with respect to the output of network 72.
With the a~ssertion of an appropriate bias oYer detector component 60, as
described in U.S. Patent No. 3,761,711 granted to Robert N. Hall on September 25,
1973 any imaging photon absorbed therewithin engenders ionization which, in turn,
creates electron-liole pairs. The charge thusly produced is collected on the ortho-
gonally disposed electrode strips by the bias voltage and such charge flows to the
corresponding node of the impedance networks 72 and 74. Further, this charge
10 divides in proportion to the admittance of each path to the virtual ground input of
the appropriate terminally disposed preamplification stage. Such charge-sensitive
preamplification stage integrates the collected charge to form a voltage pulse
proportional to that charge value. Assigning charge value designations Ql and Q2'
respectively, for the outputs 98 and 96 of the netowrk 74, and Q3 and Q4,
respectively, for the output lines 84 and 86 of network 72, the above-noted
Summation and Energy Level Derivation functions for spatial and energy data may
be designated. In this regard, the x-position of each diode defined by the orthogonal
strip geometry is found to be proportional to Q1 and Q29 and their difference i.e.
(Q~ 2), and the y-position is proportional to Q3, Q4, and their difference i.e. (Q3-
20 Q4). The energy of the incident gamma ray is proportional to Q1+Q2' and (Q3+Q4),and [(Ql+Q2) ~ (Q3+Q4)] or in the latter expression, C(Q3+Q~) ~ (Ql+Q2)] As
noted above, the operational environment of the detector array 32 and associatedamplification stages is one within the cryogenic region of temperature for purposes
of avoiding Johnson noise characteristics and the like.
As a prelude to a more detailed consideration of the spatial resolution of
gamma radiation impinging upon the entrance components of the gamma camera,
some value may be gleaned from ~n examination of more or less typical character-istics of that impinging radiation. Por instance, looking to Fig. 3 a portion of a
patient's body under investigation is portrayed schematically at 100. Within this
30 region 100 is shown a radioactively tagged region of interest 102, from which region
the decay of radiotracer releases photons which penetrate and emit from the
patient's body. These photons are then spatially selected by a portion of collimator
50 and individually detected at component 60 for ultimate participation in the
evolution of an image display. The exemplary path of seven such photons are
diagrammed in the figure, as at a-g, for purposes of illustrating this initial function
which the camera system is called upon to carry out. In this regard, the function of
--17--
~3~(~6~
collim~tor 50 is to ~ccept those photons which are traveling nearly perpendicular to
the detector, inasmuch as such emnnating rays provide true spatial image informa-
tion. These photons are revealed at ray traces, a, and, b, showing direct entry
through the collimator 50 and appropriate interaction coupled with energy exchange
within detector component 60. Photon path, C9 iS a misdirected one inasmuch as it
does not travel perpendicularly to the detector. Consequently, îor appropriate
image resolution such path represents false information which should be attenuated,
as schematically portrayed. Scattering phenomena within collimator 50 itself or the
penetration of the walls thereof allows "non-collimated" photons, i.e. ray traces, d,
10 and e, to reach the detector. Photon path trace, f, represents Compton scattering
in the patient's body. Such scattering reduces the photon energy but may so re-
direct the path direction such that the acceptance geometry of the camera,
including collimator 50, permits the photon to be accepted as image information.Inasmuch as the detector component 60 and its related electronics measure both the
spatial location and energy of each photon admitted by the collimator, the imaging
system still may reject such false information. For example, in the event of a
Compton scattering of a photon either in the patient or collimator, the energy
thereof may have been reduced sufficiently to be rejected by an energy discrimina-
tion window of the system. Photon path, g, represents a condition wherein
20 component 60 exhibits inefficient absorption characteristics such thflt the incident
photon path, while representing true information, does not interact with the
detector. As is apparent from oregoing, each of the thousands of full energy
photons which are absorbed at the detector ultimately are displayed at their
corresponding spatial location on an imaging device such as a cathode ray tube to
form an image of the source distribution within region 102 of the patient. Of coursé,
the clinical value of the gamma camera as a diagnostic implement is directly
related to the quality of ultimate image resolution.
As is revealed from the foregoing discourse, the imaging resolution of the
camera system is highly dependent upon the quality of collimation exhibited at the
30 entranee of the carnera by collimator S0. Generally, collimator 50 is of a
multichannel, parallel-hole variety, i$s performance being dictated by its funda-
mental geometric dimensions, the material with which it is formed, and the
technique of its fabrication. Referring to Figs. 4(a)-4(c), a designation of thegeometric aspects of collimator 50~ as such aspects relate to photon path travel, and
spatial intensity distribution over the corresponding spatial axis of detector com-
ponent 60 are shown schematically. Fig. 4(b) shows the photon intensity distribution
--18-
~34i ~6~$
at the mid-plane 60' of the detector due to a line source of radiation at distance B
from the collimator 50 oulwardly disposed plane defining side. Note that the source
position is designated "L". Source point, L, is located, for purposes of the instant
analysis, within a plane 104 lying parallel to the outwardly disposed plane defining
side of collimator 5Q as well as its inwardly disposed plane defining side and the
plane defined by the midpoint 60' of detector 60. The intensity distribution pattern
of photons, revealed in Fig. 4(b~, is provided under the assumption that the
collimator 50 is fixed in position. Fig. 4(a), on the other hand, assumes that the
collimator 50 moves during an exposure and produces, in consequence, a triangular
10 intensity distribution pattern of photons. A location of value "R" designates a fUll
width at half maximum (FWHM) spatial resolution. Such spatial or position
resolution capability of the camera systern may be defined utilizing several
approaches. However, for the latter designation, FWHM is derived from a
consideration that if a very small spot of radiation exits at the object plane, the
image generally will be a blurred spot with radially decreasing intensity. The
position resolution then is defined as twice the radial distance at which the intensity
is half of the center intensity.
Looking in particular to Fig. 4(c), considering the similar triangles EFG and
LMN, the resolution of collimator S0 generally may be expressed as:
Rc = DA (A + B + C) (1)
where E
A = the collimator thickness,
AE = the effective collimator thickness
due to the septal penetration,
B = the source to collimator distance,
C = the collimator to detector midplane
distance and
D = the effective diameter of each ehannel
within the multi-channel collimator
--19-
Effective diameter, D~ is considered to be the square root of the cross-
sectionfll area of a given collimator channel multiplied by 1.13.
The effective collimator thickness is given approximately by:
A = A Z (2)
where ll(E) is the attenuation coefficient of the collimator material at a photon
energy, E.
For a given collimator material, sufficiently thick septal walls are required toreduee the number of photons or gamma rays that enter within a given collirnatorchannel, penetrate the septal wall thereof and exit through an adjacent or other10 channel opening. Lookingr to Fig. 4(c3, one such gamma ray or photon path is traced
as UV . Note, that for this condition, the photon or ray passes through a collimator
vane or channel side of thickness, T, along with minimum septal distance3 W,
thereby allowing the ray or photon to exit from a channel adjacent the channel of
initial en-trance. The fraction of photons or rays traveling UY that actually
penetrate the septal wall is given by the penetration fraetion:
P = exp (- 11 (E) W). (3)
It is considered the practice of the art to design the collimator structure suchthat the penetration fraction, P, is given a value less than about 5%. In this regard,
mention may be made of the following publication:
XX. H.O. Anger, "Radioisotope Cameras," Instrumentation in Nuclear
Medieine, G.J. Hine, ed. Vol. 1, Academic Press, New York, 485-
552 ~1967).
The minimum septal distanee, W, is found from the simi-lar triangles IJK and
lJVY approximately RS: AT
W = 2D + T (4)
by assuming A is greater that 2D + T where T, as noted above, is the septal wallthickness. Solving equations (3) and (4) for the septal wall thiekness, T, gives:
T = -2D ln P
~) A + ln P (5)
30 The value, T, as set forth in equation (5) serves to define that minimal septal
thickness for eollimator S0 which is required for a given penetration fraetion, P
The geometric efficiency of the collimator is defined as the ratio of the
number of gamma rays or photons which pass through the collimator to the number
of photons or gamma rays emitted by the source. Deseribed in terms of the
eollimator para-meters, sueh effieieney may be given by:
--20--
= rKD2 12 (6)
~AE (D + T) J
where K = 0.238 for hex~gonally packed circular holes flnd 0.282 for square holes or
chambers in a square array.
AS described above, the clinical value of the gamma camera imaging system
stems importantly from the system's capability for achieving quality image resolu-
tion. Given the optimum image resolution which is practically available, it then is
desirable to provide a design which a¢hieves a highest efficiency for that resolution.
For a collimator design, it is desirable to provide a low septal penetration fraction
as well as a practical fabrication cost. Further, an inspection of equations (1) and
10 (6), given above for collimator resolution and geometric efficiency, respectively,
reveals that as resolution is enhanced, the efficiency of the collimator is diminished.
It has been determined that a multi-channel, parallel-hole collimator, the channels
of which are configured having square cross sections represents a preferred
geometric design feature. In this regard, where the latter are compared with
collimator channels formed as round holes, hexagonally packed arrays or hexagonally
packed bundles of tubes all of given identical dimensions, resolution remains
equivalent, but the efficiency of the preferred square cross sectional channel array
will be a factor of 1.4 times greater than the round hole design, while the efficiency
of the hexagonally packed bundle of tubes will be intermediate the efficiency value
20 of the above two designs. Consequently, as noted above, on the basis of maximum
efficiency at a desired resolution, the square hole cross sectional chamber design is
preferred.
Concerning the materials which may be selected for constructing the col-
limator, those e~ridencing a high density, high atomic number characteristic areappropriate for consideration. In particular, mention may be made of tungsten,
tantalum and lead for the purpose at hand. The primary criterion for the material is
that of providing a short mean free path at the photon energy level of interest. For
the desirable energy level of 140 KeV, the mean free path for photon attenuation is
0.012 inch in tungsten, 0.015 inch in tantalum and 0.016 inch in lead. Accordingly, for
30 a selection based upon a mean free path for attenuation, tungsten represents the
optimum collim~tor material. Heretofore, however, pragmatic considerations of
machinability or workability have required a dismissal of the selection of tw~gsten
and/or tantalum for collimator fabrication. For instance, for multi-channel
collimators having round channel cross sections, tungsten and tantalum are too
-21-
difficult and, consequently, too expensive for drilling procedures and, in general,
hexagon~lly packed ~rrays prov;ding such cross sections are restricted to fabrication
in le~d. Similarly, other designs formed out of the desired material do not lendthemselves to conventional machining and forrning techniques, the cost for such
fabrication being prohibitive even for the sophistic~ted camera equipment withinwhich the collimator units are intended for utilization.
In the instant preferred arrangement, a square hole collimator design, fabric-
able utilizing the optimum material tungsten, is provided. Revealed in perspective
fashion in Fig. 5, the coLlimator is shown to comprise an array of mutually parallel
10 adjacently disposed channels having sides defining a square cross section. These
channels extend to define inwardly and outwardly disposed sides which are mutually
parallel and the channels are formed axially normally to each of these side planes.
The highly desirable square structure shown in Fig. 5 is achieved utili~ing the earlier
described preferred tungsten material or tantalum, such materials normally beingdifficult or impractical to subject to more conventional manufaeturing procedures.
However, practical assembly of the collimator array 50 is achieved through the use
of a plurality of discrete rectangularly shaped sheet members, as are revealed in the
partial assembly of the collimator 114 shown in Fig. 6. Referring to that figure,
note that member 110 is formed as a flat rectangular sheet of height, h, correspond-
20 ing with desired collimator thickness, A. Formed inwardly from one edge ofmember 110 are a plurality of slots spaced in regularly recurring parallel fashion and
identified generally at 112. Slots 112 are formed having a height equivalent to h/2
and are mutually spaced to define a pitch or center-to-center spacing D + T. Theslots are formed having a width of T + e, where e will be seen to be a tolerance.
When the plurality of sheet members, for instance, as shown at 110 and 114 are
vertic~lly reversed in mutual orientation and the corresponding slots, respectively,
as at 112 and 116 are mutuaUy internested as shown, the collimator may be built-up to
desired dimensions without recourse to elaborate forming procedures. Note that the
width of slots 112 and 116 closely approximates the width of each of the sheet
30 members within the array with a controlled allowance for tolerances. In determin-
ing the value for the above described pitch of the regularly recurring slots within
the sheet members, assuming resolution criteria are met, a spacing may be selected
to match the center-to-center electrode strip spacing of a detector component 60 or
a multiple thereof so that the septal walls for the collimator 50 can be aligned with
less accurate grooves formed within the detector. E'racticPl fabrication techniques
are available for forming the slots as exemplified at 112 and 116. In particular,
1134064
chemical milling or chemiclll machining techniques are aYailable for this purpose.
With such techniques, a wax type mask is deposited over the sheets to be rnilled,
those material portions designated for remo~al being unmasked. The sheets then are
subjected to selected etchants whereupon the slots are formed. Following ap-
propriate cleaning, the sheet members then are ready for the relatively simple
assembly build-up of a completed collimator. Through the use of such chemical
milling techniques, desired tolerances in forming the slots are realizable. By
utiizing the collirnator structure shown ;n combination with optimal tungsten sheet
material, a cornputable 35 to 40 percent improvement in collimator efficiency may
10 be gained over round hole, hexagonally packed lead collimators of identical
dimension, as well as a 50 to 80 percent improvement in septal penetration
characteristics and an average 5% improvement in geometric resolution. The
collimator fabrication technique and structure are seen to offer several advantages
over more conventional collimator structures. As evidenced from the foregoing
such advantages include the availability to the design of the superior shieldingcapabilities of tungsten; a simplicity of component design and consequent ease o~
assembly and the use of optimal square hold chamber geometry for maximum
geometrical efficiency. However, to achieve optimal performance, the assembly
technique necessarily introduces small gaps at the intersections of the septal walls
20 of a completed collimator structure. These gaps exist by virtue of the tolerances
required for interlocking fit of the septal wall and the effect of gamma ray
streaming through such gaps should be considered.
In earlier commentary herein, it has been noted that septal penetration of five
percent or less of impinging gamma radiation is preferred for collimator design. It
follows, therefore, that the streaming factor for the particular collirnator structure
at hand should be assigned the same configurational parameter in the interest ofdesired unity of system design. Through utilization of a geometrlc analysis of aworst case condition, requisite lowest tolerance required for the interlocking fit of
the septal walls and for a desired source to colliFnator distance can be derived. Such
30 analysis will reveal that the slot tolerance should preferably be no more than 0.001
inch, and more preferably, should be less than that to the extent of practical milling
application.
In the discourse given heretofore concerning the function~l inter-relationships
of collimator 50 and detector array 32, no commentary was provided eoncerning the
effect of the discrete electrode strips of the detector upon ultimate image
resolution. It has been determined that, by virtue of their geometric configuration,
--23--
~3~
orthogonal strip detectors, wit~)out appropriate correction, will introduce "alias"
frequency componcnts into lhe output of the system. FoF instance, in a purely
linear system, the output of thc carnera would consist of the same spatial frequency
components as the input except with the possibility of reduced contrast. Looking to
Figs. 7(a)-~c), the aliasing phenomenon is demonstrated in connection with an
exemplary and scllematic representation of a strip electrode detector 130. In this
worst case representation, no collimator is present and the electronic resolution is
less than one strip width. Looking to Fig. 7(a), a source distribution is shown as may
be obtained, îor instance~ utilizing three discrete collimated point sources spaced at
10 equal distances of 1.5 times the strip spacing. The reciprocal of the periodic spacing
of the components depicted may be represented as, v. The source distribution shown
is one with primary frequency components of vl = 0 and v2 = 2vs/3. Such source
input is provided in the instant representation inasmuch as it combines the three
qualities which accf~ntuate an aliasing phenomenon, namely, a periodic input, 100%
contrast, and a high signal-to-noise ratio.
Fig. 7(b~ reveals a portion of strip electrode detector 130 having the earlier
described detector region grooves aligned with respect to the input signals depicted
at Fig. 7(a). The one-dimensional spatial image which may be derived, for instance,
from a multi-channel analyzer is shown in Fig. 7(c) as curve 132. By comparison, the
20 corresponding spatial image which would be received within a system incorporating
a collimator capable of resolving the input signals, a detector with strip spacing
satisfying the anti-aliasing criterion and an anti-aliasing electronic channel, is
revealed at 134. This image shows no aliased components.
Looking more particularly to the aliasing phenomenon represented at curve
132, the four lowest spatial frequency components revealed are:
(1) a component at v = 0, a zero frequency component which
represents the average value of the four peaks;
(2~ a component at v = 2v /3, which is the frequency equal to the
reciprocal of the spaciSng between one of the two outer peaks
and the average position of the two inner peaks;
(3) a component at v = vs, which is the frequency equal to the
reciprocal of the spacing between each of the four peaks; and
~4) a component at v - v /3, which is the frequency equal to the
reciprocal of the spacing between the two outer peaks.
The first two components above are the fundamental source components, while
the second two components are aliased components of the fundamental source
components centered at the first harmonic of the strip sampling frequency.
-24-
~ S fl prelllde to considering a typical rep~esentation of the spatial frequency
response of a one-clirnensional ~amma camera as revealed in Figs. 8 (a) - (d) the
modulation transfer functions (MTF) merit comment. As described in detail in
publication (III) heIeinabove~ the MTF is a measure of spatial resolution that can be
defined for linear systems and which takes into account the shape of an entire line
spread fun~tion. The reationale for such description of spatial response arises from
tlle fact that any object and its image can be described in terms of the amplitudes
and phases of their respective spatia] frequency components. The MTF is a measure
of the efficiency with which modulation or contrast at each frequency is transferred
10 by the imaging system from the object to the image. This is analogous to the
temporal frequency response of an electronic amplifier or filter. Looking now toFig. 8 (a) - 8 (d) MTF is plotted against spatial frequency, v, for a series of stages
within a gamma camera not accommodating for aliasing phenomena. In Fig. 8 ~a) a
collimator modulation transfer function (MTFC) with FWHM resolution of 1.331 is
revealed, i.e., the curve distribution, incorporating some high frequency com-
ponents, is representative of the signal passed t~ the semiconductor detector of the
camera. Fig. 8(b) reveals the output frequency spectrum of the detector which isseen by the spatial channel electronics of the camera system. An aliased frequency
spectrum is reYealed, the input signal frequency spectrum being present in the
20 output, centered at zero frequency and additional side bands of the primary input
component are present, centered at integer multiples of the strip spacing or
sampling frequency, VS = 1/1. Fig. 8(c) represents the MTF of the electronics of the
system, i.e~, the transfer function of the spatial channel electronics, while ~ig. 8(d)
shows the product of the MTF values of the curves of Figs. 8(b) and 8(c).
Accordingly, the curve of Fig. 8(d) shows the spatial frequency response of the
entire system, including the introduction of spurious spatial frequency content in the
system MTF, represented in the figure as the bump in the frequency range slightly
below vs.
Looking by comparison to Figs. 9(a) - (d) the effect of inserted correction on
30 the part of the collimator design and structure of the instant invention is revealed.
The collimator 50 design is selected to provide an MTF prefilter to limit the spatial
frequency content seen by the detector 32 to frequen~ies less than vs/2. According-
ly, Fig. 9(a) reveRls that the collimator MTF is forced to a zero value at spectrum
position vs/2. Such design insures that the fundamentul input frequency components
and the first harmonic ~requency components centered at VS do not overlap and this
~3~
condition obtains in Fig. 9(b)9 that figure revealing the alias frequency spectrum
which is processed by the electronic pickoff arrangement of the camera frorn thedetector. The spatial channel electronics complete the anti-aliasing filter system
by insuring that no spatial frequencies greater than vs/2 are passed to the imaging
system of the camera. Such post-filtering of the electronics is illustrated in Fig.
g(c). The product of MT~ conditions represented by Figs. 9(b) and 9(c) again arerepresented in Fig. 9(d) which, particularly when compared with the corresponding
Fig. 8(d), reveals the elimination of aliasing phenomena.
Turning now to the prefiltering or corrective functions carried out by the
10 collimator in controlling aliasing phe~iomena, it may be observed from the foregoing
that the system resolution of an orthogonal strip germanium detector type gamma
camera is determined by the collimator resolution, the strip width spacing, and the
resolution of the spatial channel readout electronics. The collimator is assumed to
have a Gaussian point spread function (PSF) and FWHM spatial resolution Rc. The
value of Rc should be equal to or greater than about 1.7 ( 1), where 1 is the center-
to-center strip spacing in one dimension of the detector. A more detailed discussion
of the derivation of this valus is provided in the following publication:
XXI. J.W. Steidley, et al., "The ~patial Frequency Respon~e of Orthogonal
Strip Detectors," IEEE Trans. Nuc. Sci., February, lg76.
Looking now to the specific design parameters of the collimator of the
invention, it may be recalled that collimatol resolution, Rc, has been derived
geometrically at equation (1) given hereinabove~ By now substituting the ideal
valuation, 1.7 ( 1 ) determined for antialiasin~ prefiltering on the part of thecollimaator, the ;nventive coll;mator geometry or structure may be defined.
Accordingly, the collimator is defined under the following expression:
1 7 ( 1 ) ~ ( D C) (7)
~ he collimator further can be defined utili3ing ~uation (5) above for septal
wall thickness once tlle values of the parameter of equation t7) are determined.Further9 given the value, Rc, for collimator resolution and the geometric para-
30 meters determined thereby as described above, the co~limator geometric efficiency,Q~S~ as given in equation (6) above, can be applied to further maximize theperformance of the collimator. ~dditionally, it may be noted that the suppressing
frequencies above vsJ2 input signal contributions to aliasing phenomena are ac-
commodated for.
As has been alluded to ear3ier herein, discounting entrance geometry, the
orthogonal strip position-sensitive detector is resolution limited by noise associated
--26--
~ 3~
with the detector as well as the charge d;viding network. Consequently, it is
necessary to consider the noise characteristics of the sytem from the standpoint of
minimizing the effects thereof upon resolution as well as treating such phenomena
to derive desired imaging effects. Generally, it may be concluded that the resistor
network is the dominant source of noise within the electronic spatial channel of the
system, while the resistor network, coupled with the detector leakage current,
represents the dominant noise source in the system's energy channel. As will
become rnore apparent as the instant description unfolds, spatial noise dominantly is
electrically parallel in nature, whereas energy channel noise may be considered to
10 be electrically series in nature. In the discourse to follow, noise treatment and the
like are described in conjunction with the single detector component described
heretofore in connection with Fig. 2, in the interest of clarity and simplification.
In the later portions of the instant discussion, however, the control system o the
camera will be seen to be described in conjw~ction with detector component arrayembodiments.
Noise is the random fluctuation of the preamplifier output voltage when there
is no stimulus. It is generated by imperfec$ions in the preamplifier input device,
thermal movement of charge carriers in the resistors and the bulk of the detector
and imperfections in the crystal structure of the detector. Looking to Fig. 10, an
20 equivalent noise model circuit for solid state detector components is revealed. Note
that the model reveals a detector leakage current, iD~ which is assumed to be
formed of individual electrons and holes crossing the depletion layer o~ the detector.
Such electron hole pairs are thermally generated in the depletion layer. Resistive
elements which are in parallel with the system input capacitance, CIN, generate
thermal noise which is integrated by this capacitance and appears at the pre-
amplifier input as a step function. The system input capacitance is the parallelcombination of stray capacitance at the preamplifier input and the feedback
capacitor of the preamplifier. Those resistive components which contribute to this
noise term are the high voltage bias resistor, the preamplifier feedbac3c resistor and
30 the detector bulk resistance. ~or a charge dividing resistive strip network, a portion
of the dividing resistance, ~D~ is in parallel with the detector capacitance. Since
RD is less than one hundred kiloohms, it respresents a significant noise source. The
thermal noise from resistors in series with the detector capacitance appears as a
delta function to the preamplifiers. For spec~oscopy systems, this resistance isminimized and the noise source is neglected. The noise de~reloped by the
preampli~ier input stage is modeled using a resistor, Req. Finally, a noise term
--~7--
~3~
which is not showrl in Fig. 10 is "flicker" ns)ise caused by structural changes and
surface effects in the conduction material uf the noted preamplifier input stage.
This noise aspect generally is considered to be insignificant.
Since the noise sources discussed above have a uniform power spectral density,
bandwidth limiting filtering or pulse shaping generally is considered appropriate for
maximizing the signal-to-noise ratio of the system. As suggested earlier, the
fundamental noise sources are classifiable as two types, parallel noise representing
the change due to the electron flow which is integrated by the input circuit
capacitance, and series noise represénting the charge due to the elec~on flow which
10 is not integrated by input capacitance. These noise sources are considered to be
mutually related in terms of filtering to the extent that as efforts are made todiminish one, the other increases. The high frequency component noise generally is
considered a series type while low frequency noise is considered of the parallelvariety. As has been detailed in the publications given above, the use of ~:;aussian
and the Gaussian trapezoidal noise filtering circuits has been found to optimize the
energ~v and spatial resolution values of the camera system.
Turning now to Fig. 11, a circuit model of the detector component 60 and the
resistor networks of Fig. ~ is portrayed. The discrete nature of the detector system
and the method of readout is revealed in the figure with the discrete capacitors20 forming an n x n array. Each row and column is defined by the charge measured at
the end of the resistor strings. The electron-hole pairs which are formed when agamma ray intera~ts with the detector are coLlected on opposite surfaces. A charge
enters the resistive network and flows to the terminal A or B (C or D) in relation to
the resistance between its entry point and the virtual earth terminal of each
preamplifier (Fig. 2). The intersection of the column and row defines the diode
position in which the gamma ray energy was deposited. Note in the figure, that
individual capacitances are represented which are exemplary of the inherent
capacitance of the detector itself. When considered in conjunction with the resistor
networks, as revealed in the figure, it may be noted that a particular time constant
30 or interval is required for any impinging charge to be represented by a charge flow
to the output taps of the resistor chains. Accordingly, the system must provide an
adequate time interval or time constant, TD, for this charge flow to avoid error in
information collection. In effect, it may be assumed that the detector and each of
the resistor strings of the noted impedance networks respond as a diffusive line, and
the peaking time of the preamplifier output pulses will vary as a function of the
position of interaction, xO, of an incident gamma ray. ~he voltage output of each
preamp]ifier (Fig. 2) due to the instantaneous transfer of charge Q0 at position xO is:
--28--
~3B~
V(O,x,t) = QO ~ _ xO _ ~ 2 sin ( lTL ) exp [ m l11 t~, (8)
m=l
V(L,x 9t) = Qo lXo + ~ 2 cos(m~) sin ~ L ) exp [--T--D ~(9)
m=l
where Cf is the feedback capacitance of a preamplifier in farads, L is a given linear
dimension of the detector, ~D is the time constant of the detector (i.e. ~ =
2RDCD), xo defines the position of interaction and m is a summation variable.
Examination of equation (8) and (9) show that for a time
2 (10)
i.e., an output generation time equivalerlt to one-half of the time constant of the
detector, the value of V(O,xo,t) is within 1% of its final value for all xo/L ~ .95 and
10 V(L,xo,t) is within 1% of its final value for all xJL ~ .05. Stated otherwise, the
error generated form ballistic deficit type characteristics of the system, as itrelates to the energy of one preamplifier readout diminishes to a value of 1% within
a period of one-half the time constant, TD of the detector.
By subtracting the output of the one preamplifier of a network, i.e. at the x =
L position from the corresponding amplifier output at the x = 0 position, i.e.
' (11)
V(o,xot3 - V(L~Xo~t~ = C ¦; - L ~~ m~T sin( lr ~)(1 + cos mlT) exp ~))
m=l
the following important observations may be made. Equation (11) shows that as the
spatial location of information impingement alters from 0 to L, the resulting voltage
20 readout moves from a positive unit value to a negative unit value. Stated otherwise
the output signal derived from the above signal treatment subtractive approach
ranges from + QJCf at xO-0, to -Qo/Cf a~ xO = L, mal<ing the signal twice that of
earlier suggested one preamplifier collection technique. Further, it may be
observed that the odd numbered series terms vanish, thereby reducing the position
signal peaking time. The value of equa~ion (11) is within 1% of its final value for all
values xo/I, ~. .45 nnd x/L ~.55 after a time:
-2~
~L3~
t ~ ~D (12)
Accordingly, it may be observed that through the utilizat;on of a dual preamplifier
subtractive or "antisymmetric" method of signal analysis, the necessary time
constant related signal trealmellt within the spatial channel is diminished by afactor of 4.
Turning ~ow to the conditions obtaining within the energy channel of the
system, the energy channel is derived by summing the output of each preamplifier to
obtain the voltage pulse:
V(O,xo,t) + V(L,xo~t) = C ~~ ~ n~lsin ( L ) n - cos m~) exp ( ] (13)
Note again, that the peaking time of the pulse is position dependent. At xo/L
= .5, the maximum peaking time occurs and the pulse is within 1% of its final value
at t 3 D/2. Accordingly, it may be observed that ballistic deficit or charge
collection type considerations within the energy channel will require a charge
collection period, for practical purposes, equivalent to one-half of the time constant
of the detector.
Now considering noise phenomena, as earlier discussed in combination with
ballistic deficit considerations, as derived immediately hereinabove, dominant
spatial noise, which is parallel noise, may be expressed as follows:
N = 1 ~4K~} (14)
where NqSl is the equivalent noise charge in number of electrons for one pre-
20 amplifier spatial measurements, RD is the total resistance of the resistive chain, TDis the temperature of the detector and chain, ap is a weight;ng factor of the filter, q
is the magnitude of the charge on an electron, and K is Boltzman's constant.
In the expressions given above, i.è. equations 8 through 14, the term RD is
intended as the value representing the average of the total resistance of each
resistive network. For $he exaggerated exemplary detector component shown in
Fig. 2, the term RD represents one-half the sum of the resistance v~lue of networks
72 and 74.
-30-
~13~
Note from equation (14) that the noise is proportional to the sqll~re root of the
temperature as well as tlle weighting fnctor and the time constant of the system.
As disclosed earlier, this time constant is limited by the ballistic deficit conditions
of the system. Note further that the noise is inversely proportional to total
resistance of one chain or resistor network. Therefore, it is desirable for system
efficiency to minimize the temperature ullder which it operates as well as the
weighting factor and time constant and to elevate the resistance value to the extent
practical. Equation (14) is for one preamplifier readout. Reconfiguring the equation
to represent a subtractive or antisymmetric arrangement, the following expression
10 obtains:
N SAS = 2 ~ ~1/2 (15)
From this equation, note that a subtractive arrangement permits the ballistic
deficit dictated time constant to reduce by a factor of 4, while the value of noise
increases by a factor of 2 for that same time constant. However, since a reducedtime constant (factor of 4) is involved in a subtractive arrangement, the noise value,
otherwise increased by a factor of 2, remains the same and the sign~l-to-noise ratio
is increased by a factor of 2. Recall the earlier discussion, above, that the unit
signal value runs from a positive unit to a negative unit within a subtractive system.
The value RD is difficult to increase inasmuch as a concomitant reduction in energy
20 resolution generally is witnessed for such alteration. Temperature drop can be
achieved practically, and the weighting factor, ap, can be altered to a morç or less
ideal value by appropriate selection of filtering systems. It has analytically been
determined that a 43.4 percent improvement in spatial resolution is realized if
antisymmetric summ~tion, i.~. subtractive summation, is used as opposed to the
utilization, for instance, of one preamplifier for spatial measurement.
Looking additionally to the "ballistic deficlt'l phenomenon, for thin detectors,i.e. about five mm in thickness, the detector charge collection time is small and
does not affect circuitry treating a detected signal. For thick detectors, however,
i.e. having a thickness in the range of about 2 cm, the bulk charge collection time
30 varies from approximately 100 to 200 nanoseconds. Since this collection time is
approximately the same as the collection time of the charge dividing networlc, its
contribution to ballistic deficit problems must be considered. For such systems the
optimum filtering arrangement consists OI a time invariant prefilter followed by a
-31-
~3~
gatecl integrator circuit. Such filters generally are referred to as gated integrators
or trapezoidal filters. The filter preferred îor the purpose is a Gaussian trapezoidal
filter which consists o~ a t;me invariant Gau~sian filter followed by a gated
integrator circuit. Such arrangement is revealed in more detail in the disclosure to
follow. For a detailed discourse concerning the utilization of antisymmetric
sum mation as well as the utilization of trapezoidal filtering within the spatial
channel of the system, reference is made to the following unpublished work:
XXII. Hatch, K.F., "Semiconductor Gamma Camera", Ph. D. Dissertation,
Massachusetts Institute of Technology, Cambridge, Massachusetts,
February, 197 2.
The equivalent noise charge in number of electrons for Gaussian trap-
ezoidal spatial measurements may be represented by the following expression:
~1/2
NqsGT = 2 ~kTDapTI 179 ¦ (16)
where ap is the parallel noise weighting function value for Gaussian trapezoidalsystems and TI is the integration time. Analysis of the foregoing shows that an
exeellent improvement in spatial resolution is obtained by using antisymmetric
Gaussian trapezoidal filtering. This improvement is realized because the e~fects of
"baUistic deficit" are greatly reduced.
The corresponding equivalent noise charge in number of electrons for the
20 energy channel of the system may be expressed by the following formulation:
NqE~. q ~qiDRp~o 4KrD ~1/2 (17)
An important aspect of the above energy channel and spatial channel analyses
has been observed~ In this regard, it may be recalled that opposed relationshipsstem from a consideration of parallel vs. series noise phenomena. For instance, it
has been described that energy noise is considered serial in nature whereas spatial
noise is considered to be parallel in nature. The energy noise e~uation, as shown at
(17) abovej represents a straight summation of two preamplifier outputs and the
initial parallel noise factor presented within the brackets thereof is of dismissable
magnitude. When compared with the spatial noise equation (16) above, it may be
30 observed that two separate time constant values5 IOJ Te, respectively, for spatial
resolution and energy resolution may be incvrporated within the circuitry treating
the output of the system detector. For instance, the energy resolution filtering of
the system requires a relatively extended time constant, whereas corresponding
-32-
~3~
spatifll filte~ing requires ~ relatively short one for highest signal to noise ratio
considerations. Innsmuch as the outputs of the filtering media reach the output
displays of the eamera or imaging system simultaneously, any multiple pulse errors
introd~lced into the longer time constant energy f;lter individually will be integrated
to achieve a peak value above a predesignated window function of the energy
channel (block 45, Fig. 1). Accordingly, false information generated from pulse pile-
up phenomena and the like may be rejected without recourse to more involved
discrimination circuitry. Such a desired system circuit arrangement wiLI be revealed
in the description of the control system to follow. While this description is made in
10 conjunction with the singular detector component embodiment of Fig. 2, the theory
of its operation will be seen to carry forward into the corresponding operation of the
scaled-up control system operative in conjunction with a multicomponent detectorarray.
Referring now to Fig. 12, a block schematic representation of a control system
is presented for receiving spatial coordinate outputs of the detector. In the figure,
preamplification stages 96, 98 and 80, 82 are reproduced and the outputs thereof,
respectively, are revealed at lines 110-116. Arbitrarily designating, for instance,
preamplifiers 96 and 98 as deriving energy information along an x-axis, the outputs
thereof at 110 and 112 are coupled, respectively, through lines 118 and 120 to the input
20 of a Summing and Gaussian Filtering function 122. As discussed in detail above,
function 122 operates under a relatively extended time constant, identified within
the block as Te. One output from function 122 is directed to a pulse height analysis
function 124 from along line 126. The other output of function 122 is directed to a
Gate Control and Display Control function 128 from along line 130. This is an energy
derivative pulse, as identified at line 130, and provides a start pulse input to function
128. Output lines 110 ~ld 112 also provide the spatial channel input to Antisymmetric
Summation and Gaussian Filtering function 132. From function 132~ a subtractive
filtered signal is directed along line 134, to a C;ated integrator 136 operating under
an integrating period correspondir.g with time constant ~O. Control into the gated
30 integrator, for instance, establishing the time constant value, To, emanates from
gate control function 128 through line 138. Additionally, a reset control is provided
to the integrator from line 140.
Similar to the x-axis spatial chanel inputs, the y-axis spatial channeL inputs
deriving through lines 114 and 116 are introduced into an Antisymmetric Summation
and Gaussian Filtering function shown at block 142. The output from block 142, as is
present at line 144, is introduced to a Gated Integrator function 146, structured
--33--
identically to Gated Integrator function 136. Time constant ~0 control over
integrator 146 is asserted from gate control function 128 through line 148, while reset
control is asserted from line 150. The outputs from the x-axis Gated Integrator
Function 136 is presented along line 152 to a Photographic Record readout 154 and
through lines 152 ~nd 156 to a Persistant Display Scope 158 may be utilized for
purposes of patient positioning and other information desired by the operator.
Similarly, the y-axis spatial channel in~ormation derived from Gated Integrator
function 146 is presented along line 160 to Photographic Record output 154 and
through lines 1~0 and 162 to Persistent Display Scope readout 158. Readout control
10 to Photographie Record 154 and Persistent Display Scope function 158 is derived
from Gate Control and Display Control function 128 through lines 164 and 166. The
control asserted thereby is one wherein outputs 154 and 158 are not actuated or are
blanked until control function 128 receives an input display signal from Pulse Height
Analysis function 124 through line 168. Interrogation of function 128 is provided from
control 128 through line 170. Inasmuch as a relatively extended time constant, Te, is
utilized at Summing function 122, any pulse pile-up phenomena will be integrated to
derive a peak pulse level beyond the upper window limita~ions of the channel
analyzer operating within function 124. Accordingly, error otherwise introduced into
the system from the spatial channels is blanked upon the assertion of an interroga-
2~ tion request from line 170 and a responding blanking type signal or no response signalfrom function 124 $hrough line 168.
Looking now to Fig. 13, a block schematic diagram is provided showing the
basic components of the Gated Integrator and associated functions depicted
generally at blocks 136 and 146 in Fig. 12. Note that the circuit includes an input
amplifier 172 which feeds, in turn, into a delay line 174. Delay line 174 is utilized to
insure that the integrator gates are open before any spatial informational pulsearrives thereat. ~he circuit further includes a base line restorer, as at 176, which
opeates in cooperation with gated integra~or 178. The output of integrator 17~ is
directed to an output amplifier 180, the output from which is directed along lines 152
30 or 160, as shown in Fig. 12, depending upon the particular orthogonal sense of the
incoming signal.
A corresponding and more detailed schematic representation fo the circuit is
revealed in Fig. 14. Referring to that figllre, either of the coordinate spatial inputs
as developed at lines 144 or 134 (Fig. 1~ is asserted through an input resistor 182 to
an amplification stage 184. Stage 184, corresponding to amplifier block 172 in Fig.
13, includes a feedback line incorporatin~ ~eedback resistor 1869 as well as a ground
refe~erice input at line 188. Delay line 174 is shown represented at 190 receiving an
input from olltput 192 of amplifier 184. A resistor 194 is coupled between delay line
190 and ground, whil~ the output thel~eof is AC coupled through capacitor 196 to the
input of ~ base line restorer f~mction. The base line restorer is of a Robinson type
as is generally described, for instance, in the following publication:
XXIII. Robinson, L.B., "Relations of Baseline Shift in Pwlse Amplitude
Measuremen$s", Rev. Sci. Inst., 32, 1961, p. 1057.
Essentially9 the restorer function is provided for the purpose of assuring a netzero charge value at the gated integrator input prior to the reception of any input
10 signal. Further, the restorer defines the maximum charge that can be placec] on the
coupling capacitor 196. In the absence of the restoring function, the gated
integrator would integrate areas below the baseline as well as under the Gaussian
shaped spatial sign~l. For carrying out its assigned functions, the restorer includes
an emitter-follower stage at NPN transistor 198, the base of which is coupled
through resistor 200 and line 202 to one side of capacitor 196. The emitter of
transistor 19~ is coupled through a resistor 204 to -Vcc potential, while its collector
is coupled through a resistor 206 to +Vcc. The restorer function additionally
includes a current supply network oeprating such that, upon the occurrence of
spurious elevations of current, accommodation is made to ¢ontrol the quiescent
20 point at the emitter-follower stage 198. Note that this current supply includes a
PNP transistor 208, the emitter and base of which, respectively, are coupled throug
resistors 210 and 212 to +Vcc. This base, additic,nally, is coupled to ground through a
resistor 214. The collector of transistor 208 is coupled through diode 216 to line 202
Rnd through diodes 218 and 220 to a variable resistor 222, the termini of which are
connected between the positive and negative sides of the supply voltage.
The output of the base line restorer function is coupled through resistor 224 toone terminal, for instance the source, of a field effect transistor (FET) 226
representing the intput of the gated integrator function, while the opposite
electrode of the transistor is coupled to line 228~ Line 228, in turn, is directed to
30 one side of an integrating amplifier 230. The gate input to FET 226 is present at
line 232 and is shown as selectively receiving a signal designated y from the
control function 128 (Fig. 12). Also influencing line 228 is a network including line
234 and variable cQpacitor 236 which is coupled to receive an input designated y.
The opposite input to amplification stage 230 is coupled to ground through line 238.
Amplification stage 230 performs an integrating function by virtue of its feedback
~3~
connection with an integrating capac;tor 240 coupled between lines 2~2 and 244. A
shunt;ng res;stor 24~ ;s coupled between lines 242 and 244 in parallel with capacitor
240 and is selectively activnted by a reset gate present as field effect transis-tor
(FET~ 248, the so~lrce and dr~in terminals of which are connected in switch defining
fashion within line 244 and the gate input to wh;ch at l;ne 250 is configured toselect;vely rece;ve a reset signal ;dentified as, f3, from Gate Control Function 128
(Fig. 12). A variable resistor 252 is connected between the positive supply voltage
and the interconnection of resistor 246 with FET 248. The output of amplification
stage 230 is present at l;ne 254 and is coupled through a variable capacitor 256 and
10 line input 258 for selectively rece;ving a signal input identified as, B.
The output at line 254 of the gated integrator is directed through resistor 260
to the input of a unity gain inverting amplifier 262 which includes a feedback line
incorporting resistor 264 and is connected with ground reference at line 266. The
output of the amplifier, at line 268, is that represented in Fig. 12 either at line 1S2
or line 160 and is directed to the readout components of the camera system. As will
be apparent ;n the discussion to follow, gate control over integrator 178 is derived by
the noted signal inputs into lines 232, 234 and 250 and 258.
Looking to Figs. 15 and 16, the control circuit represented in Fig. 12 at 128 isdisclosed in more detail in combination with a timing sequence diagram. At time, t
20 = 0, as shown in the timing diagram of Fig. 16, the sytem is prepared to process an
- incoming set of signals or pulses. The time derivative of the energy pulse or signal
dE/dt is directed along line 130 to a comparator ~92. When its value e~ceeds a
reference voltage representing the lower level of the window level established by
evaluation or Pulse Height Analysis function l24 (Fig. 12) it serves as a start or to
actuate the control system. The voltage reference against which the derivative o~
the ener~y pulse or signal is compared is inserted from line 290 to the comparator.
These predetermined, preliminary signal level conditions being met, comp~rator 292
provides a positive going output pulse at line 294 which is introduced to a dual, D-
type flip-flop FF-l. ConventionaIly, the D form of flip-flop incorpor~tes an
30 actuating (clock) inpUt signal $erminal, Ck, along with a signal input terminal, D.
The flip-flop output signal Q becomes 1 at the time of a 0-to-1 change at the clock
terminal. In conventional manner~ the Q output of the flip-flop represents the
inverse OI the Q output. The D flip-flop also is characterized in incorporating a
clear feature designated at "Cl" in the diagram. To further facilitate the
description of the circuit, Boolean designation is utilized to represent input or
output values. For instance, a "low" signal is considered to be one having a potential
--36-
~L3~
essentially at ground nnd is typically represented by a logical "zero". Conversely, a
"high" signal is considered positive and may be depieted by a logical "one".
Returning to Figs. 15 and 16, with the presence of a positive going pulse at line
294, f]ip-flop FF-l is clocked such that its Q output at l;ne 298 assumes a high value
and its Q output at line 296 a~sumes a low value. Note that the Q output of flip-flop
FF-l is identified as B and is introduced to the reset gate of each integrator, as
shown in Fig. 12-14. With the opening, for instance, of reset gate transistor FET 250,
the shunt about timing capacitor 240 is removed to enable the integrating amplifier.
Similarly, the Q output of fli~flop FF-l assumes a low status and, by connection10 through line 296, couples the ~3 signal input to the gated integrator as at line 258 in
Fig. 14. This ~ signal output of the flip-flop FF-l is used to compensate for charge
injection into the feedback capacitor caused by the capacitance coupling betweenthe gate and drain electrodes of F~T 248.
The Q output of flip-flop FF-l additionally is presented through line 298 to theB input terminal of a monostable multivibrator M-4, and3 through line 300, to the B
input terminal of monostable multivibrator M-l. Accordingly, these multivibrators
are triggered, the Q output of multivibrator M-4 being programmed for elosing each
integrator input gate for a time slightly greater than the base width of the Gaussian
shaped spatial pulses. In this regard, note that the Q output, as is represented at
20 line 232 of the miltivibrator M-4, carries a y signal which is introduced into the
input gate of FET 226 (Fig. 14). Simultaneously, an inverted y input is providedalong line 234 to variable capacitor 236 to provide compensation for off charge
;njection. The gated integrator then commences an integrating mode of perform-
ance, the time over which operation is controlled by multivibrator M-4. It rnay be
observed that multivibrator M-4 retains this output state in correspondence with a
spatial time constant determined interval, ts, aæ is more clearly protrayed in Fig. 16.
Note in that figure, the representation of a Gaussian spatial pulse, S, corresponding
with the activation of the integrator function.
As noted above, the Q output of flip-flop FF-l also is introduced to the B input30 terminal of monostable multivibrator M-l. With the presence of the forward edge of
this signal at line 300, the Q output of the latter alters from a low to a high value
and retains such value over an interval, te, selected for delaying the start of the
display sequence until the energy pulse has been analyzed at pulse height analysis
function 124 as shown in Fig. 12. Note that this interval, te, always will be greater
than the integrating interval, ts. The Q output of multivibrator M-l is coupled--
through line 302 to the A input of monostable multivibrutor M-2. On the occurrence
-37-
~3~
of the negative edge of the pulse of the Q output o~ multivibrator M-l, multivibrator
2 is triggered and the resultant Q output transition thereof is directed along lines
304 and 306 to driver circuit Dl. The output at line 170 of driver Dl serves as the
earl;er described interrogation pulse directed to the pulse height flnalysis function
12~ described in connection with Fig. 12. Note aclditionally, that line 306 extends to
one input of a NAND gate Nl. Acco~dingly, the signal from line 306 is inverted and
introduced through line 308 to the A input terminal of monostable multivibrator M-
5. This input serves to enable the latter to permit the carrying out of a full control
cycle. During typical display operation of the camera system, the 'linterrupted'l and
10 llon/offll inputs to NAND gate Nl are high at the option of the operator. By
converting either or both to a low value, multivibrator M-5 is inhibited to, in turn,
inhibit the displays as referred to earlier in Fig. 12 at 154 and 158.
The remaining components of the circuit function on the basis of whether an
interrogating signal issued from line 170 to Pulse Height Analysis Function 124 (Fig.
12) has been responded to, along line 168, to indicate a pass or no pass condition of
signal energy level. If function 124 does not respond to the interrogating pulse from
line 170, thus indicating that the peak value of the energy pulse did not fall within
the window setting of the evaluation function, multivibrator M-5 receives no signal
input at terminal B thereof. Additionally, upon the occurrence of the negative edge
20 of the Q output signal of multivibrator M-2, multivibrator M-3 is triggered from line
304 such that its ~ output at line 310 asserts a clearing signal through AND gate ANI
to the clear input terminal, Cl, of flip-flop FF-l. The output thereof, as reflected at
the multivibrator M-4, causes the integrator to be reset. With this operation, the
sytem is short cycled, and the through-put rate thereof advantageously is increased.
Note, that the opposite input st line 314 of AND gate ANl is normally high by virtue
of its connection through line 316 and resistor 31X to a positive voltage source.
Assuming that multivibrator M-5 has been en~bled from the line 308~ A, inpu~
thereto ~nd that a positive response has been received from Pulse Height Analysis
function 124 and line l68 at the B terminul input thereto, the multivibra~or will react
30 by developing a positive output pulse at its Q terminal and line 320, while a pulse of
opposite sense is developed at its Q output along line 332. The Q output signal at
line 320 is directed to line 32~ whereupon it addresses the elock input, Ck, of D flip-
flop FF-2. In consequence, the Q output o~ flip-flop FF-2, at line 322, converts to a
low value which is asserted at the B input of multivibrator M-3 to inhibit the output
thereof. The short cycle feature thereby is inhibited. This signal at line 324 may
also be utilized ~or clocking a scaler or count recording apparatus through a driver
circuit D2.
-38--
~3~
The outputs of multivibrator M-5 also are utilized to switch a Z-axis driving
ci~cuit from a negative to positive voltage, thereby turning on the display functions
represented in Fig. 12 at 154 and 158. In this regard, note that line 324, carrying the
Q output signal of multivibrntor M-5, is connected to the gate electrode of a field
effect transistor (FET) 328. By corresponding connection, the Q output of
m~lltivibrator M-5 is asserted ~long lines 33~ and 334 to the ~ate input field effect
transistor (FET) 3300 Note that the drain-to-source channel of FET 328 is connected
through resistors 336 and 338 to a positive voltage source, while the correspondin~
source-to-drain channels of FET 330 are coupled through resistors 340 and 342 to a
10 negative voltage supply. The respective opposite sides of FET's 328 and 330 are
connected through line 344 and line 346 to one input of the Z-Axis amplifier 348 and
are biased such that, under conditions wherein monostable multivibrator M-S, is not
clocked, the output of amplifier 348 is retained at a negative value. Upon the
clocking of multivibrator M-5, however, FRT 330, in effect, closes while FET 328opens, to cause the output of amplifier 348 to change from a negative to positive
value, thereby permitting the activation of display and record functions 154 and 158
- (Fig. 12).
Zener diodes as at Zl and Z2 are present in the input network to Z-axis
amplifier 348 for the conventional purpose of voltage regulation, the diodes being
2D commonly coupled to ground at their respective anodes. Additionally, the re
spective cathodes of the diodes 71 and Z2 are coupled at the common connections of
resistors 336 and 338, 340 and 342. I,ooking further to the outputs OI multi-
vibrator M-5 as they respond to an energy eva]Luation input at line 168, the positive
edges of the output signal at Q thereof also activates a multivibrator M-6 in
consequence of the connection of line 332 with the B terminal thereof. Multi-
vibrator M-6 serves to provide a delay function assuring an adequate interv~l for
turning off the electron means of display scopes and the like. The Q output of
multivibrator M-6 is coupled through line 352 to the B terminal input of anothermonostable multivibrator M-~. The positive going edge of the Q output signal of the
30 multivibrator M-6 triggers multivibrator M-7 to provide a corresponding pulse signal
at its Q output at line 354. Lnie 354 is coupled through ANl~ gate AN2, the output
of which is coupled through line 356 to the clear terminal, Cl, of flip-flop FF~2. The
opposite input to AND gate AN-2 is operator preselected andis asserted from line314. With the presence of a clearing input ~o flip-flop FF-2, the Q output thereof at
line 322 returns to a high status which, in turn, is imposed at the B input of
multivibrator M-3 which, in turn, functions to clear flip-flop FF-l by virtue of the
--39--
~11 34L~
connection therewith of its Q output at line 310 through AND gate AN-l. With theclearing of flip-flop FF-I, the gated integrator is discharged or reset through the
input signfll at line 250, described earlier in connection with ~ig. 14.
With the noted return of monostable multivibrator M-3 to its standby state,
the control system is fully reset and ready to proce~ss another set of information
signals. If the output of comparator 292 is high at this time, the system will not
process such incoming information. This lockout feature prevents the partial
integration of spatial pulses too narrowly spaced in insertion time. Note that
comparator 292 is coupled to receive the ~ signal output from the Q terminal of
flip-flop ~F-l through lines 296 and 297. This input signal is utilized by the
comparator as a block to any enabling of the system to respond to incoming sign~ls
until such time as a full cycle of evaluation has terminated. Fig. 16 reveals the
time-based correspondence between the output of comparator 292 and the Q output
or B signal of flip-flop FF-l. In the absence of such B signal input from line 297,
error would be introduced into the system, for instance, by virtue of the generation
of start signals at line 294, integrator timing is disrupted to invalidate an ongoing
signal processing procedure. As may be evidenced from Fig. 16, the ~ signal input
from line 297 (Elip-flop FF~ terminal) serves to inhibit comparator 292 until the
reset point of a given signal processing cycle.
As noted earlier, any display of spatial pulses which overlap is prevented
because, for the optimized filtering system, the base width of the spatial Gaussian
pulse is less than the peaking time of the energy Gaussian pulse or interval of energy
analysis. Because of this, should two or more garnma rays photoelectrically interact
with the detector and their totnl energy be absorbed in a time less than the rise-
time of the filtered energy pulse, the resulting energy pulse peak value would not
fall within the window defined at pulse height analysis function 124. As a
consequence, the control system would carry out a short cycle function as revealed
in the timing diagram of Fig. 16 by a dashed line alteration of the curves.
Examination of the dashed curves of ~ig. 16 re-veals that upon interrogation of
30 Pulse Height Analysis 124, should no response signal be received therefrom within
the interrogation interval defined by multivibrator M-2, the negative going edge of
the output thereof causes multivibrator M-3 to carry out a reset function, thereby
inhibiting the carrying out of the remainder of the signal processing cycle.
As noted earlier, it is important that the detector function of the gamma
camera be capable of accepting photon information from as broad a spatial reg;on as
possible. Inasmuch as the size of detector crystals inherently is limited by the
--~0--
~.3~
technigues of theil fabrication, it becomes necessary to conjoin a p]urality of such
detector components in some manner wherein a broader region of radiation may be
witne~ssed and imaged.
One preferred technique for associating the discrete detector components
provides for their mutual operational intercomlection in subgroupings oî pre-
determined mlmbers9 for instance, 4, which subgroupings then are coupled with the
control system of the camera. An array of detector components having a requisitecamera entrance area acceptance geornetry then is formed preferably of a sym-
rnetric compilation of component subgroupings. One practical size for the detector
10 array comprises four subgroupings each of which is formed of four detec tor
components. With such an array, the control system advantageously may operate byobserving the performance of the subgroupings as they are represented in quad-
rature. Another aspect to be considered in "scaling-up" the camera system for
clinical utility resides in the earlier-discussed feature permitting their accepting
and properly imaging information derived from two discrete photon energy levels.Accordingly, the scaled-up gamma camera would incorporate a control system
accommodating both of these desired features. In the discourse to follow, the
general signal treatment described hereinabove in connection with Figs. 12--16
remains substantially the same with an appropriate multiplication of functions
20 where necessary to accommodate for the greater number of generated signal inputs
from the detector groupings.
Turning to Fig. 17, a composite detector, formed as an array of discrete
detector components, is revealed generally at 360. This subgrouping of four
detector components, as identified at 362, 364, 366, and 368, may, for instance~ be
combined with three additional subgroupings to form a full detector array compris-
ing four subgroupings incorporating A total of 16 detector components. Of course, a
gre~ter or smaller number of detector components may be combined to form an
array of desired dimension. In the interest of clarity, only one such quadrant
designated subarray, as at 360, is described in conjunction with a control system.
30 Detectors 362-366 are dimensioned having mutually equivalent areas designated for
the acceptance of impinging gamma radiation. Such equivalency serves to achieve
accurate ultimate image read-out from the camera system. The detector com-
ponents 362-366 are of the earlier-described orthogonal strip array variety, each
strip thereof being defined by grooves~ Note in this regard, that detector
component 362 is formed having strips 370a-370d located at its upwardly disposed
surface and defined by ~rrooves cut intermediate adjoining ones of these said strips.
The opposite face of the detector component 362, similarly, is formed having strips
372a-372d defined by intermediately disposed grooves arranged orthogonally with
respect to the grooves at the upper surface. Detector component 364 is identically
fashioned, having strips 374a-374d at its upwardly disposed surface, each being
defined by intermediately disposed grooves; the lower surface of the detector being
Iormed having orthogonally disposed strips 376a-376d defined by intermediately
disposed grooves. The corresponding strip arrays of detector component 366 are
shown to comprise identically disposed strip groupings as at 378a-378d and 380a-
10 380d. Similarly, detector component 368 is shown to be formed of identicallystructured mutually orthogonally disposed strip arrays 382a-382d and 384a-384d.
Components 362-368 are illustrated expanded from one another for purposes of
illustration only, it being understood that in an operational embodiment these
components are internested together in as practical a manner as possible. To
achieve an informational spatial and energy output from the diserete detector
components, the strip arrays each are mutually associted along common coordinatedirections. This association is carried out between components 362 and 364 by leads
386a-386d, coupling respective strips 374a-374d of component 364 with strips 370a-
370d of component 362. In similar, parallel coordinate fashion, leads 388a-388d are
20 provided for connecting respective strips 382a-3$2d of component 368 with strips
378a-378d of component 366.
The outputs of the thus mutually coupled strip arrays of the upwardly disposed
faces of the detector components are coupled with an impedance network, repre-
sented generally at 390. Network 390 is configured comprising serially inter-
connected discrete resistors 392a-392i. Interconnection between respective strips
370a-370d and points intermediate resistors 392e-392i is provided by leads 394a-394d, while corresponding interconnection between strips 378a-378d with the
intermediate connections of resistors 392a~392d is provided by leads 396a-396d.
In similar fashion, the arrayed strips 372a-372d at the lower surface of
30 component 362 are coupled with respect to strips 380a-380d of component 366 by
leads 398a-398d. Similarly, strips 376a-376d at the lower face of component 364 are
respectively coupled with corresponding strips 384a-384d of component 368 by leads
400a-4QOd. The thus associated strip arrays of the lower faces of the detector
components are connected with the second impedance network, identified generallyat 402, in similar fashion as the orthogonally disposed upward surfaces. Note~ for
instance, that strips 380a-380d of the lower surface of component 366 are connected
--42-
~3gL~
to intermediate respective discrete resistors ~04a-~lO~e of network 402 by l~ads40Ba-406d. Simiiflrly, strips 384a-384d of the lower surface of component 368 are
connected with respective discrete resistors 404f-404i of network 402 through leads
408a-408d. Thus interconnected, the four discrete detector components provide
spatial coordinate parameter outputs; i.e. x-designated coordinate outputs at lines
410 and 412, which are identified thereat as (xlA) and (xlB). In like manner, the
spatial coordinate parameter outputs of the lower surfaces of the detector com-
ponents are present at] nes 414 and 416 and are y-designated, being labeled in the
drawing, respectively, a (ylA) and (ylB).
Fig. 18 reveals a first output treating arrangement, present as one set of
filtering and control electronics which operatcs in conjunction with the quadrant
detector array of Fig. 17. In that figure, spatial coordinate parameter outputs, or
x-designated coordinate outputs (xlA), (xlB) and (ylA), (ylB) are represented,
respectively, at lines 410-416. These outputs, as at lines 410-416, are shown arranged
to address discrete preamplification stages respectively revealed at 430-436. In this
regard, note that the output at line 438 of preamplification stage 430 is introduced
to an x-Channel Antisymmetric Summation Gaussian Filtering and Gated Integrator
function, shown at 440, while the corresponding input from preamplification stage
432 is directed along line 442 to that same function. The summing and filtering
20 functions at 440 operate on the x-coordinate outputs introduced thereto in the same
fashion as described above in connection with Fig~s. 12-16. For instance, the inputs
from the x-spatial coordinate outputs are subtractively summed and, following
appropriate filtering and pulse shaping as by noted series of integrations and the
like, an output from block 440 is provided as an x-designated coordinate channelsignals at line 446.
The outputs of x-channel amplification stages 430 and 432 also are directed,
respectively, through lines 450 and 444 to Summing and Gaussian Filtering function
448. As described in conjunction with Pigs. 1~-16, function 448 includes an initial
stage deriving the time derivative of the summed energy signal provided from lines
30 444 and 450 and submits such derivati~!e signal, from along line 452, to a Gate
Control and Start Logic function, identified at block 454. Such signal evidencing a
predetermined requisite level to provide a preliminary assurance of valid spatial
information, the start logic function of block 454 responds to provide gate control
over Filtering and Summation Function 440, as through line 456.
The corresponding y-coordinate outputs of amplification stages 434 and 4369
respectively, are coupled through lines 45B and 460 to a y Channel Antisymmetric
-43-
Sumrnation {md Gallssian l~iltering f-mction 462. Configured in similar fashion as
function 440, the signals introduced to function 462 ar subtractively summed,
appropriately filtered, and pulseshaped by a series of integrations to provide a y-
designated coordinate channel signal at line 464. Control over the gated integration
function, as well as filtering at block 462 is provided from gate control and start
logic block 45~ ~s thr ough line 466. In fashion similar to that described in
connection with Figs. 12-16, the control system further includes an Energy dis-
crirninator, revealed at block 470, which reeeives the summed energy signal output
at line 472 from Summing and Gaussian Filtering function 448. As before, Energy
10 r)iscriminator 470 provides a pulse height analysis of the energy signal to evolve an
accurate evaluation thereof as to the presence or absence of valid image informa-
tion. Upon interrogation thereof through line 474 from gate control function 4~4,
and response thereto at line 476, the signal treating cycle is permitted to continue.
However, as described earlier, where the pertinent energy signals fail to meet the
window criteria of Energy Discriminator 470, gate control 454 will effect a
resetting of the summing functions, as from lines 478, 480, and 482 to carry out the
earlier-described short-cycle operation, thereby permitting the system to more
rapidly and efficiently process a next incoming spatial signal. It may be noted that
Energy Discriminator function 470 may operate effectively within the system even20 though more than one photon energy level of information is asserted. Recall that it
is desirable to accommodate the system to the utilization of more than one
radiopharmaceutical9 each such radio-labeled substance having a different gamma
ray energy characteristic. Because the gemanium detector system of the inventionenjoys a considerably improved resolution characteristc, the discriminator ~7 0 is
capable of performing its assigned function in a practical fashion at this stage of the
control system. In this regard, the germanium detector e2~hibits a capability of from
3 to 4 keV resolution range as opposed to a generally observed range o~ about 15 keV
achieved with more conventional scintillation type cameras. Accordingly, Energy
Discriminator 470 readily may be adjusted to pass those energy signals representing
30 the lower acceptable level of the lower photon energy designated radiopharma- ceutical.
In accordance with the invention, the filtering and control electronics for a
given quadrant also incorporates a Peak Detector function represented at 484.
Detector 484 is coupled through lines 472 and 486 to receive the energy signal
generated from summing function 448. The detector 484 serves to hold the peak
value of this signal7 thereby providing an analogue storage function to accommodate
-44-
for variations in signal treatment times AS are represented for instance, between
~ntisymmetric Surnmation functions 440, 462 and the energy additive Summing
function at 448. Detector 48d~ iS a.ssociated in time control fashion with gate
control 454 th~ough line 490 and may be reset therefrom through lines 478 and 480.
The peak value output of detector 484 is presented along line 492 to an Energy
Channel Driver 494 for ultimate presentment to quadrant processing control
circuitry described later herein. Note that the energy channel signal present flt line
496a is designated, Qle.
With the occurrence of an appropriate acceptance of the validity of a spatial
10 signal at Ener~ Discriminator 470, the x-designated coordinate channel or spatial
signal at output line 446 iS delivered to an x-Channel Driver 498, the output ofwhich is present in line 500a. Note that this channel signal is designated ~lx
Similarly, the y-designated channel signal, having been treated at function 462 flnd
admitted to the system by the Energy Discriminator 470 and gate control functions,
is presented nlong line 464 to a y-Channel Driver 502, the output OI which is present
at line 50~a and identified as, Qly.
The information now delivered from each quadrant of the overall imaged
region now, for purposes of convenience, is designated by the noted labels: Qlx' Qly
flnd Qle. ln the immediate discussion to follow, the composite detector arrfly is
20 assumed to be functioning in qufldrflture and, thereby, developing corresponding
signals from four distinct multi-component quadrflnts. These quadrants are repre-
sented by a "Q" with the noted subscripts altered by the values 1-4. Gate control
454 also provides a clocking or datfl acceptanc~3 signal seq~encing input to thecontrol system at line 506a the signal from which is designated, Ql' and is arranged
to receive a reset signal, designated, Qlr~ as at line 508a. The latter signal is
selectively derived from the quadrant processing control system to be described in
conjunction with ~ig. 19.
Turning to that figure, the noted processing arrangement is revealed in block
schematic fashion and is shown to include three multiplexing input networks, an x-
30 Position Multiplexer 510, a y-Position Multiplexer 512, and an Energy Multiplexer
514. The inputs to multiplexers 510-514 derive from each of the four quadrant
circuits and, as an example of the designations utilized in the instant figure, for the
quadrant circuit described in connection with Fig. 18, such inputs are represented
and labeled at lines 500a, 5n~a, 4~6a, 506a, and 508a. Correspondingly, the inputs
from the three remaining circuits to eaeh~of the multiplexers are represented and
labeled, respectivley, at 500 ~d, 504 ~d, 496 ~d. AdditionaUy, the outputs from
--45--
Gate Control ~nd StaIt logic functions, as described, for example, at 454 in Fig. 18
are represented, respectively at 506a-so6d as leading to F.I.F.O. Memory 516, while
the input to functions as at 454 in Fig. 18 are represented as output ~ines 508a-508d
extending from Reset Drive function 520. Note, additionally, that the quadrant
signals leading to the quadrant processing control arrangement of Fig. 19 are
identified by an ascending nurneration suffix for each input function, i.e. the inputs
for x-Position Multiplexer 510 are identified as Qlx-Q4x' the inputs to y-Position
Multiplexer 512 are identified as Ql -Q4, the inputs to Energy Multiplexer 514 are
identified as Qle~Q4e5 the data acceptance s;gnal inputs to F.I.F.O. Memory are
10 identified as Ql-Q4 and the outputs of Reset Drive function 520 are identified as
Q1R Q4R-
In conventional fashion, multiplexers 510-514 perform a switching type function
wherein the channel signals addressed thereto are selected and forwarded into the
system upon appropriate control logic commands present as coded actuating signals.
These multiplexers are regulated from a quadrant interface control function,
represented within a dashed line boundury 522. Function 522, in addition to
incorporating the F.I.F.O. Memory and input clocking 516 and Reset drive function
520, includes a 4-to-3 Line Decoder 524 and a Sequential Control function 526.
F.I.F.O. (first-in, first-out) Memory and Input Clocking 516 is conventionally formed
20 incorporating somewhat independent input and output stages or networks. It serves
within the system as a derandomizer which receives and collects or records the
randomly generated data acceptance signals, which are presented at lines 506a-
506d. These quadrant labeled signals are received and serialized at the input
clocking stage of F.I.F.O. Memory and Input Clocking 516 following which, an
appropriate signaling or clocking pulse is sent to the Sequential Control 526, through
line 5?2 which tells the sequential control that quadrant information is available. In
consort, the quadrant identification signal is passed through output 534 and 570 to
the 4-to-3 line decoder 524 and the reset drive 520. The resultant coded actuating
signals are presented to the multiplexers along grouped liens 536, 537, and 539
3D which, in turn, signal the appropriate gates within respective multiplexers 510, 512,
and 514 to pass the retained spatial and energy signals to a Sample and Hold
Amplified function (S/H). In this regard9 note that the output of x-Position
Multiplexer 510 is provided along line 542 to a Sample and Hold Amplifier 544, while
the corresponding y-Position Multiplexer 512 output is presented along line 546 to
Sample and Hold Amplifier 548. Similarly, the output of Energy Multiplexer 514 is
provided along line 550 to Sample and Hold Amplifier 552. Sample and Hold
--46-
~3~
~mplifiers 544, 548 and 552 are utilized within the circuit as an analog storagemedium so that the aforementioned quadlant circuits can be reset for processing
incoming signals. Line 530 extends through line 556 to Sample and Hold Amplifier552; through line 558 to Sample and Hold Amplifier 548; and directly to Sample and
Hold Amplifier 544. A command from Sequential Control 526 emanating from line
530 pl'OVideS an initial sample command whereupon the multiplexers output signals
are sumpled by the Sample and Hold Amplifiers. Following a select interval, a hold
signal is passed to amplifiers s~L4-ss2, following which a next succeeding interval is
provded. In consort with the issuance of the hold signal to amplifiers 544-552, a
10 reset command signal is passed by the Sequential Control through line 566 to the
Reset Drive circuit, 520. Since the spatial and energy information contained in the
quadrant circuit being addressed is now stored in the processing circuit, the
appropriate quadrant circuits can be reset through the appropriate reset line Q1R-
Q4R (lines 508a-508d). In carrying out the latter functions, it may be noted that
reset drive function 520 derives the quadrant selective information through lines 570
from the F.I.F.O. Mernory 516, at the end of the reset command a clockout pulse is
sent to the F.I.F.O. Memory 516 along line 532 from the Sequential Control unit 526.
By doing this the information at the output of F.I.F.O. Memory 516 is clocked out so
that the next usable information appears at F.I.F.O. Memory 516 output. The fact20 that valid information appears at F.I.F.O. Memory 516 output is sent along line 572
to the Sequential Control circuit 526 for current or future use. At the time valid
ouput appears at F.I.F.O. Memory 516 output the 4/3 Line Decoder 524 decodes it
for processing by the multiplexer circuits. During the latter interval, the energy
signal, now passed to the hold functoin of amplifier 552, is present at the output
thereof at line 560, during which period it further is introduced at line 561 and
analyzed by a Two-Channel Analyzer 562. Analyzer 562 provides pulse-height
analysis reguisite to evaluating the energy levels of the earlier noted two radio~
pharmaceuticals, for example, which may be utilized within the system. Note thatthe analysis performed at function 562 is the second within the system, the initial
30 evaluation being carried out in the earlier-described circuitry associated with each
quadrant of the detector. Should the energy sign~l passed to the Two Channel
Analyzer 562 fail to meet the window criteria for either select photon energy level,
an appropriate representation or signal is provided at output line 564 to Sequential
Control function 526. If a valid, information present pulse was received from
~.I.F.O. Memory and Input Clocking 516 along line 572, then teh sample and hold
amplifier 54d~, 548, and 552 are placed in the sample mode and the process described
--47--
~3~
above repeats itself. If a val;d information present pulse was not received, theSequential Control 526 waits until one is received before the process described
above repeats itself. Where more than two photon energy levels are selected for the
system, an appropriate multichannel analyzer is substituted at component 562.
An advantageous aspect of the inVeIltiOn resides in the controlled inter-
relationship between multiplexers 510-514 and corresponding Sample and Hold
Amplifiers 544-552. As controlled by sequential control 526, an initial clockinginput to multiplexers 510-514 causes a quadrant signal received thereby, in given
arrival order, to be clocked to the appropriate ones of Sample and ~Iold Amplifiers
10 544, 548, and 552. FoUowing the initial interval described hereinabove, commencing
with the noted hold function, the initial treating system may be cleared in
anticipation of a next quadrant signal to be processed as described above. This
feature advantageously improves the throughput rate of the overall system permitting correspondingly improved imaging performance. With the clearing of the intial
treatment or input networks as well as clocking OI F.I.F.O. Memory 516, the
entrance portion of the quadrant processing circuitry is preapred to accept the next
succeeding quantum of information form the quadrant circuits. Upon appropriate
command from se~uential control 526 following an interval suited to permit the
noted two-channel analysis to be carried out at block 562, the x- and y- spatial20 output signals of Sample and Hold Amplifiers 544 and 548 respectively which are
passed along lines 580 and 582 to Divider networks 584 and 586 are stable and
proportioned respectively to the x- and y- positions. The latter networks serve the
function of normalizing the spatial signals received from lines 580 and 582 withrespect to the particular photon energy of the detector interaction which they
represent. The corresponding energy signal introduced at this point to the system
and analyzed at 562 may be representd as, Qe~ while the spatial signal may be
represented ss, ll~e. From the earlier discussion presented herein, the spatialinformation, cL, which the system, notwithstanding quadrature data, provides as
spatial visual information may be represented by the expression:
(~ L ), (18)
- 48 -
~,3~
where, L, is equivalent detector lenyth and, xO, is the position of interaction of a
gamma ray. The spatial information quantity, ~, also may be derived and expressed
by the relationship:
E ) (19)
Equation (19) reve~ls the function of Diviclers 584 and 586, i.e. by dividing by QE~
the measured energy channel signal, as it r epresents one or more photon energy
levels, the desired spatial signal, tY, is derived. In carrying out this function, the
divisor of the last expression in Equation (19) is derived as a signal from Sample and
Hold Amplifier 552 through lines 560 and 588. Line 588 is coupled with a Bias
Amplifier 590 which provides an appropriate d.c. offset to prevent the presentment
of a nonzero denominator to the divider circuits. The outputs of Bias Amplifier 590
are shown directed along line 592 to Divider 586 and ~long line 594 to Divider 584.
Because of the advantageous high quality resolution of germanium type solid state
detectors, the dividing function provided herein is required only for systems
intended to utilize radioisotope imaging sources to present more than one photonenergy level but can be used with a monoenergetic source.
The thus normalized x- and y- spatial channel signals are directed, respec-
tively, from Dividers 584 and 586 along line 596 and 598 to an x-,y- Display Bias and
Orientation Control function, identified at control block 600. At function 600, the
signals introduced thereto are weighted in correspondence with the quadrant or
coordinate from which they weere derived by virtue of a weighting signal input from
control 526 as directed thereto through line 602. Control 600 also may include
appropriate circuitry providing for an operator se]lection of the alignment of the x-
and y-axes wherein they may be interchanged for desired clinical analysis purposes.
Control 600 is coupIed in information exchange relationship for x-channel informa-
tion with a Camera Display Control 604 through lines 606 snd, for y-chflnnel
communication purposes through line 608 with the same control. Display activations
information to the Camera Display Control 604 is derived from the output of
Sequential Control 526 through lines 666, 610 and 612. In conventional fashion, the
spatial data and the display on/off control is coupled to a CRT Display 614 through
the three channel input lines extending thereto and identified at 616, 618 and 620.
In similar fashion, a Patient and Positioning Display function is provided by
Positioning Oisplay Control circuit 622 operating in conjunction with a positioning
display readout, i.e. CRT tube or the I~ke, as represented at block 624. Spatial
--49 -
~39L~
channel information is dh~ected to Positioning l~isplay Gontrol 622 from lines 626
and 628 which, in turn, are connected with respective l;nes 606 and 608. A display
activating signal is sent to controi 622 from Sequential Control 526 through lines
666, 610, 612 and 630. The spatial and energy channel intercoupling between control
function 622 and its corresponding display readout at 624 are shown represented by
lines 632, 634, and 636.
The control system also incorporates a readout identified in Fig. 19 as a
Spectrum and Energy Window Display represented at block 638. This readout servesto permit the operator of the system to adjust the window settings of the two-
10 channel analyzer 562 to achieve accurate evaluation of the energy level of the
particular isotopes utili~ed for imaging. The infs~rmational input of function fi38 is
derived from the energy related output level of Sample and Hold Amplifier 552, as is
present at line 560. To convert this level to a corresponding transitional signal, a
Linear Gate 640 is arranged to receive the energy signal at ]ine 560 and transmit a-
corresponding transitional type signal along lines 642 and 644 to a Multi-channel
Analyzer 646. The channel outputs of analyzer 646 are presented along lines 648-654 to a spectrum display control 6S6 which, in turn, provides display control over
Spectrum and Energy Window Display 638 by virtue of its connections therewith
represented at lines 658-662.
The system also provides a count rate meter 664 which is coupled with th0
display control signal from Sequential Control 526 through lines 666 and 670. Such a
device apprises the operator of the presence of imaging photons at the region ofanalysis and their reaction rate with the system. Note, additionally, that the output
display control of Sequential Control 526 also is directed through line 666 and 670 to
Linear Gate 640. The output of Linear Gate 640, at line 642, also is directed to a
Total Count function 672, the latter serving to apprise the operator of the quantity
of imaging information received at the camera. Other components which might be
incorporated within the system, which are not shown in the figure, may include~ for
instance a total time recorder apprising the operator of the span or interval of30 operation of the camera for given clinical analysis as well as an isotope gain control
for regulating the Summation and Filtering function described hereinbefore in
connection with Fig. lB.
With the final imaging of a given energy signal, at CRT Display 614, sequential
control 526 directs a signal alcng line 530 to the Sample and Hold Amplifiers 544-
548 and 552 placing them in the sampling mode. Since, as described above, the
information from the next quadrant is presented as each Sample and Hold Amplifier
-5~-
`` 3l~3~
input through lines 542, 546 ~nd 55() an improvement in system throughput rate is
realized.
In Figs. 20-22 as are described hereinafter, another form of composite
detector is revealed which provides a "row-colunn" form of readout of the spatial
and energy data within a select grouping, n, of detector components. In each of
these embodiments shown, a reduced component linear dimension over which
resolution is required is achieved to improve the resolution of the entire system.
Two embodiments of this "row-column" arrAngement are revealed wherein a larger
effective detector area is provided while the earlier described time constant, ~D~ is
10 minimized to improve the response rate of the system to processing interaction
generated image signals.
Referring now to Fig. 20, another composite detector formed as an array of
discrete detector components is revealed generally at 680. Illustrated in exploded
fashion, the detec$or 680 is comprised of a plurality of detector components, four of
which are shown at 682, 684, 686, and 688. Components 68~-688 are dimensioned
having mutually equivalent areas as are intended for acceptance of impinging
radiation. This required equivalency serves to achieve an accura$e ultimate image
readout from the camera system. In the absence of such equivalency, distortion at
such readout, exhibiting a discontimlity of image information, would result. The20 detector components illustrated are of the earlier-described orthogonal strip array
variety, each strip thereof being àefined by grooves. Note~ in this regard, thatdetector component 682 is formed having strips 690a-690d located at its upward
surface and defined by grooves cut intermediate adjoining ones of the said strips.
The opposite face of detector component 682 similarly is formed having strips 692a-
692d defined by intermediutely disposed grooves arranged orthogonally with respect
to the grooves at the upper surfac~. Detector component 684 is identically
fashioned, having strips 694a-694d at its upwardly disposed surface and lower
surface, orthogonally disposed strips 96a-96d each strip being defined by inter-mediately formed grooves. Similarly, detector component 686 is formed having
30 strips 698a-698d at its upward surface defined by intermediately disposed grooves,
while its lower surface similarly is formed having strips 700a-700d defined by
intermediately disposed grooves arranged orthogonally with respect to the grooves
OI the upward surface. Detector component 688 may be observed having strips
702a-702d at its upward surface defined by intermediately designated grooves, while
its lower surface is formed with strips 704a-704d separated by intermediately
disposed grooves arranged orthogonally to the grooves of the upward surface of the
~omponen$.
Detector components 682-688 as well as similar cornponents in later figures
are illllstrated expanded from one another for purposes of illustration only, it being
understood that in an operational embodiment these components are internested
together in as p~actical a manner as possible. To achieve an informational spatial
~nd energy output from the discrete detector components, which essentially is
equivalent to that output which would be realized from a large detector of
equivalent size, the strip arrays are functionally associated under a geometry which,
as noted above, may be designated "row" and "column" in nature. In this regard,
note that an impedRnce network, shown generally at 707, is associated with the
10 strips 694a-694d of detector component 684. This network incorporaltes discrete
resistors 706a-706e which are tapped at their common junctions by leads 708a-708d
extending, respectively, to strips 694a-694d Thus configured, network 707 closely
resembles the impedance networks described herein in connection with Fig. 2. Note
however, that output lines 710 and 712 of network 707 extend to and are coupled in
parallel circuit relationship with the corresponding output of a similar impedance
network, identified generally at 714. Network 714 incorporates discrete resistors
716a-716e which are tapped at their common interconnections by leads 718a-718d.
Leads 718a-718d, in turn, respectively extend to strips 690a-69ûd of detector
component 682. Accordingly, the upwardly disposed surfaces of detector com-
20 ponents 682 and 684 are identically associated with respective impedance networks714 and 707, while the latter are interconnected in row fashion and in parallel circuit
relaticnship to extend to principal output terminals, as are depicted generally at 720
and 722. It may be noted that the information collected at these principal terminals
represents one imaging spatial coordinate parameter of a select directoinal sense
i.e. along a designated row.
Looking now to the fundamental interrelationship o~ detector components 686
and 688, a similar coordinate parameter direction or row-type informational
collection network is revealed. In this regard, note that the impedance network,shown generally at 724, is configured comprising discrete resistors 726a-726e, the
30 points of common interconnection of which are coupled with respective leads 728a-
728d. Leads 728a-728d, in turn, respectively, are connected with strips 698a-698d
at the upwardly disposed surface of detector component 686. Likewise, an
impedance network, shown generally at 730 incorporting discrete resistors 732a-
732e, is associated with detector component 688 by leads 734a-734d extending,
respectively, from strips 7D2a-702d to the points of common interconnection of
network discrete resistors 732a-732e. Additionally, the output lines as at 736 and
~5~-
738 of network 730 are connected in pa~allel circ~lit relationship with the output of
network 724 to provide row readout termini, respectively, at 740 and 742. Here
again, a row-type ~irectional spatial coordinate parameter is provided at the
upwardly disposed surface of the composite detector 680.
Looking now to the lower surfaces of the detector components, it may be
observed that the orthogonally disposed strips of detector component 682 are
associated with an impedance network identiîied generally at 744. Network 7~aS
incorporfltes discrete resistors 746a-746e which are coupled from their mutual
interconnections by leads 748a-748d, respectively, to strips 692a-692d of detector
10 ~82. Similarly, the orthogonally disposed strips of detector component 686 are
associated with an impedance network 750. In this regard, network 750 is formed of
discrete resistors 752a-752e which, in turn, are coupled, respectively, with strips
700a-7~0d by leads 754a-754d. Tlle output of impedance network 744 is connected
by leads 756 and 75~ to the corresponding output of impedance network 750 to
provide column direc~ional coordinate parameter outputs, as at 760 and 762, which
serve to collect all spatial information of the associated paired surfaces of
detectors 682 and 686.
Looking to the lower surface of detector component 684, note that a network,
designated generally at 764, incorporating discrete resistors 766a-766e is function-
20 s~lly associated with strips 696a-696d, respectively~ by leads 768a-768d.
In similar fashion, an impedance netowrk, designated generally at 7709 is
associated with the orthogonally disposed strips 704a-704d at the lower surface of
detector component 688. Note that the networh:, incorporating discrete resistors772a-772e, is functionally associated with the array of strips 704a~704d, respectiv~
ly, by leads 774a-774d. Networks 764 and 770 are eleetrically eoupled in parallel
circuit fashion by collector leads 176 and 778 and extend to principal collection
points of termini 780 and 782. Thus interconnected, the lower surfaces of detectors
684 and 688 are coupled in column readout fashion to provide another spatial
coordinate parameter of direction parallel with the correspondin~ lower surface
30 strip array readout arrangement of detector components 682 and 686.
With the row and column readout intercoupling of the detector components as
shown in the figure, it may be observed that the capacitance exhibited by all
discrete detector components, taken together, remains the same as if only a single
detector were operating within a camera. AccordinglyJ the signal treating circuitry
and logic of the camera, advantageously, may be designed to accommodate for the
charge collection time constant of a single detector. Connection with the row and
-53-
column reac3Outs for given spatial coordillate parameters outputs will be seen to be
provided by treating Cil'CllitS which distribute coordinate channel spatial and energy
channel signals ;nto anflly~ing and distributing circuitry. Such circuitry is described
in more detail in connection with Eigs. 19 and 23-25. Preamplification stages, as
described in connection with Fig. 2, are coupled with each row readout point as at
720 and 722 or 7~0 and 744, as well as with each column readout, as at 760 and 762
and 780 and 782. Such preamplification stages generally are located within or near
the cryogenic environment of the detector itself. The molmting of the contact leads
between each of the networks and an associated strip array surface of a detector10 generally may be carried out by resort to biased contact configurations.
The composite detector arrangement or interrelated detector component
mosaic also may be formed utilizing detector structures which incorporate surface
disposed resistive layers to achieve spatially proportioned charge readout character-
istics. Such a detector composite is revealed generally in Fig. 21 at 800. Referring
to ~hat figure, the composite detector, or portion thereof, 800, is shown to comprise
four discrete detector components 802-80~. The opposed surfacers of the detectorcomponents, which are situated generally normally to impinging radiation, are
formed having a resistive character. This resistance is provided, for instance, by so
lightly doping the n-type surface as to achieve a region of resistive character, while,
20 similarly, so lightly doping the opposite surface with a ~type acceptor as to achieve
a surface resistive character thereat. The readout from these resistive surfaces is
collected by conductive strips which, for the case of detector component 802, are
shown on the upward surface at 810 and 812 and at the lower surface at 814 and 816.
Conductive surfaces 810 and 816 may be deposited upon the detector component ~02,
for instance? by conventional evaporation techniques utilizing a highly conductive
metal such as a noble metal, i.e. gold.
Concerning the techniques for developing the noted resistive regional char-
acter within the surfaces of detector components 802-8087 mention may be made ofthe following publications:
X~IV.Owen, R.P., Awcock, M.L., "One and Two Dimensional Position
Sènsing Serniconductor Detectors," IEEE, Trans. Nucl. Sci., Vol.
N.S. -15, June 1968, Page 290.
XXV.Rerninger, W.H.9 "Pulse Optical and Electron Beam Excitation o~
Silicon Position Sensitive Detectors", IEEE, Trans. Nucl. Sci., Vol.
V.S. 21, P~ge 374.
With the impingement of r~diation upon detector component 802 and resultant
development of an interaction therewithin, charge will be collected on the opposed
--54--
~39L~
surfaces, as discussed above, and will split proportionally at the impedance defined
surfaces and collect at the conductive strips 810~816. For the upwardly disposedsurface, these charges then are collected ~long conduit 818, coupled with conductive
strip 812, and conduit 819, coupled wi~h conductive strip 810. The adjacently
disposed detector 804 is fashioned in similar manner, the upward surface thereofincorporating a resistive surface layer or region formed in cooperation with
conductive strips 8~0 and 822~ The lower surface of detector component 804 is
formed incorporating a similar resistive layer or region functionally associated with
conductive strips 824 and 826. Note that the latter conductive strips are arranged
10 orthogonally with respect to those at 820 and 822. Conductive strip 820 is coupled
by a lead or conduit 828 to cGnductive strip 812 of the detector component 802,
while conductive strip 822 is coupled by lead or conduit 830 to conductive strip 810
of detector 802. Thus interconnected, it will be apparent that any interaction
occurring within detector component 804 will be "seen" as a charge division between
strips 820 and 822, for one coordinate parameter, along leads 828 and 830, as well as
output conduits 818 and 819. As is apparent, a desirable simplification of the
structure OI the composite detector is available with this form of row readout.
Looking to the adjacently disposed row of detector components 8û6 and 808, it
may be noted that detector component 806 is formed incorporating resistive la~ers
20 or regions in its opposed surfaces aligned for the acceptance of radiation and,
additionally, incorporated conductive strips as at 832 and 834 at the extremities of
its upward surface as well as orthogonally oriented conductive strips 836 and 838
about the extremities of its lowermost and oppositely disposed surfaces.
Identically structured detector component 808, similarly, is formed having
resistive surfaces or regions arranged normally to the direction of radiation
impingement The surfaces also incorporate conductive strips, as at 840 and 842 at
the upwardl~y disposed side and, at ~44 and 846, orthogonally disposed at the
lowermost surface.
Coupled in similar rowtype fashion as detectors 802 and 804, the condu~tive
30 strips of deteetors 806 and 808 are directly electrically associated by leads 848 and
850. Note, in this regard, that lead 848 extends between conductive strips 840 and
832 while lead 850 ex$ends between conductive strips 842 and 834. The output of
that particular row at the upward surface of the composite detector is represented
by leads 852 and 854.
A columnar interconnection of the detector components is provided between
the orthogonally disposed conductive strips 814 and 816 of detector 802, respectively,
as by leads û56 ancl 858, to similarly disposed conductive strips 836 and 838 ofdetector 806. The columnar readouts for the paired detector components are
present at conduits 860 and 862 extendi~g, respectively, from conductive strips 836
and 838.
In similar fashion, the columnar association of detector components 804 and
808 is provided by leads 864 and 866 which, respectively, extend between conductive
strips 824 and 826 of detector 804 to corresponding conductive strips 844 and 846 of
detector component 80~. The r eadouts for the column association of detectors 804
and 808 are provided by conduits 868 and 870 extending, respectively, from
lû conductive strips 844 and 846 of detector component 808.
As in the embodiment of Fig. 20, the output conduits 818, 819 and 852, 854 are
of a '1row" variety having fl designated spatial coordinate parameter and are
addressed to initial preamplification stages prior to their association with logic
circuitry for deriving imaging information for that particular spatial coordinate.
Similarly, the "columnar" outputs at conduits 860, 862 and 868, 870 are directed to
preamplification stages, thence to appropriate circuitry for treating that spatial
coordinate parameter. It will be understood, of course, that the number of detector
components formed within a matrix or array thereof depends upon the fisld of view
desired for a particular camera application as well as the practicalities for
20 retaining such components under appropriate cryogenic temperature conditions
during operation.
The foregoing examination of the composite detector structures, represented
in Figs. 20 and 21 reveals certain consistent charaeteristics between the embodi-
ments. For instance, as alluded to above, the effective areas presented to radiation
impingement of the discrete detector components must be substantially equivalent,
in order to avoid distortion in an ultimately developed image. Additionally9 these
components should be as closely nested as possible and aligned such that the spatial
coordinate which may be designated for each surface evoles what has been termed
as a "row-column" orientation. In the latter regard, an observation of this geometry
30 shows that the leads interconnecting the impedance networks or the impedance
stucture i.e. at the surface region of the detector components~ connect them
directly, whether in the parallel-series connection of the embodiment of Fig. 20 or
the interconnection of conductive strips shown in Fig. 21. Another aspect typifying
the structure of the invention, reveals that any two adjacent surfaces of any two
adjacent detector components exhibit spatial coordiante parameters of a common
directional sense and, more particularly, two adjacent of the coplanar surfaces of
--56--
ally two acljucent dctector components are dispo.sed within a linearly oriented
grouping arranged to exhibit a common spatial coordinate parameter directional
sense. Because the composite detector embodiments shown in Figs. 20 and 21
operate substantially in the same functional manner, their outputs are identified
with the sarne spatial coordinate directional labels. For instance, the x- designated
coordinate outputs at lines 722 and 720 of the embodiment of Fig. 20, respectively,
are identified as ~XlA) and (XlB); while the parallel row y- designated coordinate
outputs as at lines 742 and 740, respectively, are identified as (X2A) and (X2B).
Similarly, the orthogonally disposed y- designated coordinate parameter outputs, as
10 represented for instance, at lines 76~ and 760, respectively, are identified as (YlA~
and (YlB). Ncxt adjacent to that column of the composite detector, are the
detectors whose outputs are represented Rt 780 and 782 and are identified,
respectively, as ~Y2A) an (Y2B). This same labeling procedure will be seen to beutilized in the composite detector emhodiment of Fig. 21.
An important aspect of the "row-column" interconnection of the discrete
detector components resides in the realization of an effective reduc tion in that
detector linear dimension over which resolution is evaluated. More specifically, an
improvement is experienced in the resolution of the camera system which may be
expressed by the equation:
~x = ~E L
E ~20~
Where, ~x, represents spatial resolution in terms of distance; ~E, is absolute energy
resolution; L, is length of Q detector component as measured parallel to the
directional sense of an associated impedance network; and E, represents the energy
of an incident photon interacting with the detector. Within the right hand side of
equation (20) above, the expression, ~E~ is readily identified as the fraction (or
percentage) of energy resolution and is fixed for a given input energy. Accordingly
any increase in the value of, L, directly and adversely affeets the spatial resolution.
Where the detector components are not interconnected by the "row-column"
technique, the value, L, in the expression above becomes larger. For example if the
30 detectors pictured in Fig. 22 were connected as a single detector, the measuring
distance would be ~L, effecting a doubling of the noted spatial resolution value to
the detriment of final imaging. Another feature characteristic of the detector
"row-column" interconnection resides in- the presence of a common detector
component for each combination of an associated row and column. Stated
--57--
otherwise, a row or column configuration also may be designated as an orthogonally
disposed linearly oriented groupin~ of charge collecting surfaces. Any interaction
within ~ny given common component will provide ~- and y-designated coordinate
output signals from the thus associated linear surface groupings.
A third embodiment for "row-column" interconnection of detector components
exhibiting this spatial resolution advantage is revealed in Fig. 22. Referring to that
figure, a composite detector formed as an array of discrete detector components is
revealed generally at 880. As in the earlier-discussed embodiments, detector or
detector portion 880 is sho~n in exploded fashion for purposes of clarity and
10 comprises a plurality of detector components four of which are shown at 882, 884,
888, and 886. Components 882-888 are dimensioned having mutually equivalent
areas as are intended for acceptance of impinging radiatoin and are formed as of an
orthogonal strip array variety, each strip thereo being defined by grooves formed
within the detector surfaces. Of course, other, strip-defining configurations wiLI
occur to those skilled in the art. Detector 882 is formed having strips 890a-890d
defined by grooves cut within its upward charge collecting surface. The oppositeface of detector component 882 similarly is formed having strips 892a-892d defined
by intermediately positioned grooves arranged orthogonally with respect to the
grooves at the upper surface. Detector cornponent 884 is identically fashioned,
20 having strips 894a-894d formed at its upwardly disposed charge collecting surface;
and at its lower surface, orthogonally disposed strips 896a-896d, adjacent said strips
being defined by intermediately formed grooves. Similarly, detector component 886
is formed having strips 898a-898d at its upward sllrface, adjacent ones of the strips
being defined by intermediately disposed grooves, while its lower surface similarly is
formed having strips 900a-9OOd defined by interm~ediately disposed grooves arranged
orthogonally with respect to the grooves of the upward surface. Detector component
888 may ~e observed to have s~ips 902a-902d at its upward surface adjacent ones of
which are defined by intermediately designated grooves, while its lower surface is
formed with adjacently disposed strips 904a-904d separated by intermediately
30 disposed grooves arranged orthogonally to the grooves of the upward surface
thereof.
In the instant embodiment, strips 894a-894d of detector component 884 are
directly, electrically associated with corresponding row strips 890a-890d of com-
ponent 882 by elecb ical leads, respectively identified at 906a-906d Note, that no
impedance network is interposed intermediate the strip groupings as in the earlier
embodiments. However, an impedance network, designated gener~lly at 908, is
--58--
~3~
u~soci~ted with the termini of strips 890a-890d opposite the edges thereof coupled
with electrical leads 906a-906d. Network 908 comprises serially associated discrete
resistors 910a-9lOe which are tapped at their common junctions by leads 912a-912d
extending, respectively~ to strips 890a-8gOd. The output, or readout points for the
thus defined '~ow" of the composite detector assernbly are represented at 914 and
916 and are provided lhe same respective spati~ or x- designated coordinate
parameter output labeling, (xlB), (xlA) as are present in the corresponding "row" of
the embodiments of Figs. 20 and 21.
The corresponding upwardly disposed surfaces of components 886 and 888 are
10 connected in similar ashion. For instance, strips 902a-902d are electrically coupled
with strips 898a 898d by respective electrical leads 918a-918d. The "row" coupling
thus provided is associated with an impedance network shown generally at 920.
Network 920 is formed comprising serially associated discrete resistors 922a-922e
which are tapped at their common interconnections by leads 924a-924d. Leads
924a-g24d, respectively, extend to strips 898a-898d of detector 886. The principal
termini of the thus defined "row" are identified at 926 and 928, having outputs
respectively labeled ~x2B), (x2A).
Looking now to the lower surfaces of the detector components, the ortho-
gonally disposed strips of detector component 882 are electrically coupled as shown
20 with the corresponding strips of detector component 886 by electrical leads 930a-
930d. The thus coupled strip arrays of those detector components are associated in
"columnar" fashion with an impedance network identified generally at 932. Network
932 comprises serially associated discrete resistors 934a-934e, the interconnections
between which are connected as shown with strips 900a-9OOd of component 886 by
leads respectively identified at 936a-936d. The re~dout termini for the thus defined
"column" association o~ detectors 886 and 882 are present at g38 and 940 and thecorresponding spatial or y- designàted coordinate parameter outputs are identified
respectively as (ylA) and (ylB).
The lower surfaces of detector components 884 and 888 similarly are asso-
30 ciated in "columnar" readout fashion, strips 896a 896d of the former being electric-
ally connected through respective leads 940a-940d to strips 904a-904d of the latter.
The thus established 1'columnar" readout is associated with an impedance networkidentified generally at 942 and comprising serially associated discrete resistors
944a-944e. Strips 904a-904dJ respectively, are coupled with the interconnection of
the resistors 944a-944e of network 942 by leads 946a-946d. As in the earlier
embodiments, the principal readouts of the thus defined "columnar" detector
-59-
component coupling are represented at 948 and 950 and their spatial coordinate
parameter outputs are labeled, respectively, (y2A) and (y~13). From the foregoing
description of the composite detector arrangement 880 it may be observed that the
row-column association of the components thereof enjoys the noted spatial resolu-
tion advantages, however, the time constant characteristic thereof will reflect a
higher capacity evaluation.
Figs. 23 and 24 reveal filtering and control electronics which operate in
conjunction with the quadrant of composite detectors arrayed in the "row-column"manner described hereinabove in connection with Figs. 20-22. Note that the spatial
10 coordinate parameter outputs from the arrays shown in those figures aredesignated
(xlA), (xlB) and (x2A), (x2B) for the row readouts and (ylA), ~ylB) and (y2A), (y2B) for
the corresponding columnar readouts. Looking to Fig. 23, the x-channel spatial
coordinate parameter outputs as are derived from one such x-channel row type
readout, to wit (xlA), (XlB) are again reproduced at intput lines 1000 and 1002,representing the input addressing respective discrete preamplification stages 1004
and 1006. It should be understood that each row within each quadrant would
incorporate the initial or first signal treating functions, including the control
electronics revealed in Fig. 23 and that the components to be described in
connection therewith are substantially identical in function as those described
20 heretofore under substantially the same labeling, the description now being reduced
to single channel analysis in the interest of clarity and simplicity. The output at
line 1008 of preamplification stage 1004 is introduced to an x-Channel Anti-
symmetric Summation and Trapezoidal Filtering function 1010, while the correspond-
ing input from preamplification stage 1008 is directed through lines 1012 and 1014 to
that same function. The Summing and Trapezoidal function at 1010 operates on thexl-Channel Spatial signals introduced thereto in the same fashion as described a~ove
in connection with ~igs. 12-16. For instance, the inputs from the xl-Channel aresubtractively summed and, foL~owing appropriate Trapezoidal Filtering and Gaussian
Shaping, for instance~ by the noted series of integrations or the like, an output from
30 function block 1010 is provided at line 1016.
The outputs OI amplification stages 1004 and 1006 also are directed, respect-
ively, through lines 1008, 1018 and 1012, and 1014, to the Summing and Gaussian
Filtering function 1020. As described earlier in detail in conjunction with Figs. 12-16,
function 1020 includes an initial stage deriving the time derivative of the summed
energy signal provided through lines 1014 and 1018 and submits such derivative signal,
from along line 1022, to a Control function depicted generally at 1024. When this
-60-
~39L~
signal evidenees a predetermined requis;te level to provide a preliminary assurance
of valid spatial information, A start logic function within block 1024 responds to
provide gate control over the F iltering and Summation function at block 1010.
Controls over gates and the like of function 1010 are asserted, for instance, from
lines as at lp26, while appropriate information feedback is derived from the
Filtering and Summation function 1010 from communication line 1028. Control 1024~lso communic~tes with an Energy Discriminator function 1030 which receives the
summed energy signal output of block 1020 from along lines 1032 and 1034. As in the
earlier embodiments, Energy Discriminator 1030 provides a pulse-height analysis of
10 the energy sigrlal deriving from Summing function 1020 for purposes of evolving an
initial evaluation thereof as to the presence or absence of valid image information.
For instance, discriminator 1030 evaluates the energy signals in correspondence with
the lowest photon energy level to be accepted from those radioisotopes extant
within the noted region of clinical interest. Appropriate acceptance or rejection of
the energy level is signalled along line 1036 to Control function 1024 andJ in the
absence of an appropriate such level, the latter function serves to reset the system
by carrying out the earlier-described short-cycle operation. As in the earlier
embodiments, the circuit of Fig. 23 further includes a Peak Detector function 1038.
Associated with Control 1024 through line 1040 and receiving the summed or energy
20 signal from block 1020 through line 1032, Peak Detector 1038 serves to hold the peak
value of the signal passed thereto to provide an analogue storage function for
accommodating variations in signal treatment at times as are representedJ for
instance, between ~ntisymmetric Summation and Trapezoidal Filtering function 1010
and Summing and Gaussian Filtering function 1020. The peak value output of
detector 1038 is presented along line 1042 to an energy channel driver 1044 for
ultimate presentment to quadrant processing control circuitry from along line 1046.
Note th~t the energy channel signal at that line is identified as, Qle.
The output at line 1016 of Antisymemtric Summation and Trapez~oidal Filtering
function 1010 is presented to an xl-Channel driver 1048 for delivery to quadrant30 processing control circuitry through line 1050. Note, as in the earlier embodiments,
this coordinate channel signal is identified by the label Qlx Similar-ly, the output
of control block 1024 is provided at line 1052 and the coordinate channel signalthereat is identified by the label l'x-Channel Ql " Additionally, for purposes of
effecting a full cycle or short cycle termination, an input provided at line 1054 for
carrying such appropriate signal is labeled "x-Channel Q1R"-
,
~&
Looking to Fig. 24~ the corresponding initial input treating or columnar or y
Channel processing circuit is revealed, it again being understood that this c;rcuitrepresents that associated with only one column within a "row-column" detector
component array, similar such circuits being required for each such column. Of
course, the term "row'l or "column" is for descriptive purposes only and designates
one given coordinate parameter of directional readout from a detector matrix or
mosaic. Looking to the figure, column readouts (ylA), (ylB) are asserted, re-
spectively at input lines 1060 and 1062 of input preamplification stages 1064 and 1066.
The output of amplification stage 1064 is directed through line 1068 to a yl-Channel
10 Antisymmetric Summation and Trapezoidal Filtering function 1070 performing the
operations described hereinabove. Similarly, the output of preamplification stage
1066 is introduced through lines 1072 and 1074 to signal treatment at function 1070,
the yl-Channel signals being subtractively summed, appropriately filtered and pulse
shaped by a series of integrations to provide a y-Channel signal at line 1076. This
signal is introduced through a yl-Channel driver 1078 from which it exits at line 1080
for introduction as a signal designated "~lY" to second or further treatment at a
processing control function.
The yl-Channel signals also are introduced from lines 108 2 and 107 2 to a
Summing and Gaussian Filtering function 1084 which additively sums and filters the
20 signals to generate an energy signal which is submitted, as through line 1086, to an
Energy Discriminator function 1088. As before, Energy Discriminator 1088 carriesout a pulse height analysis of the energy signal to provide an accurate evaluatio
thereof as to the presence or absence of valid image information. The lower value
selected for this analysis corresponds with the acceptable lower value for the lowest
photon energy selected for receipt and treatment by the system. The output of
discriminator function 1088 is directed through line 1090 to Control function 1092. In
addition to providing appropriate gate control over Summation and Filtering
function 1070 through line 1094~ Control fun~tion 1092 also re~eives the time
derivative of the yl-Channel energy signal from ~long line lOS6. As in the earlier
30 embodiments, this signal generally is obtained from an initial stage within Summing
and Gaussian ~iltering operations perforrned at block 1084. This derivative signal
serves both to provide an appropriate start signal logic for the Control 1092, as
described earlier herein and, additionally, will be seen to provide a coincidence
signal for later control over the entire multi-component detector arrangement ofthe invention. The output of control function 1092 is present at line 1098 and is
identified, for illustrative purposes, as "y-Channel Ql". This signal is introduced to
--62--
~13~
the noted second or further treatment at a processing control function, that same
circuitry providing a reset input to contro~ 1092 identified flS "y-Channel Q1E~" and
submitted from along line 1100.
Turning now to Figs. l9 and 25, the second treating or quadrant processing
control arran~ement for use with tbe "row-column" readout arrangements for
detector arrays remains substantially similar to the system described above in
connection with Fig. l9. However, in consequence of the isolated readout geometry
necessarily present in a "row-column" interCOnneCtiOrl, R further arrangement isrequired to properly identify and treat a data pair input deriving from the given x-
10 and y- Channel outputs of any given detector component within the array.
Accordingly, the quadrant interface control function represented in Fig. 19 within
dashed boundary 522 is replaced by the slightly revised arrangement at 522' in Fig.
25. For purposes of clarity, the latter drawing incorporates broadened arrows torepresent a multi-line input as would be derived from the multiple channels of
networks of Figs. 23 and 24 as well as the multiple line couplings already depicted
in l?ig. 19. Further in this regard, to represent the several inputs from each row of
the detector array, typical ones of which are represented in Fig. 23 as 'tx-Channel
Ql"' and in Fig. 24 as "y-Channel Ql"' the corresponding multiple channel input for
such readouts is îdentified as l'x-Channel Qnll and y-Channel Qnl'. Additionally, the
20 reset signals from the processing circuitry at Fig. 25 are generally denoted by the
labels llx-Channel QnR" and "y-Channel QnR". 1`hese labels represent typical reset
signals, two of which are identified above in connection with the description
associated with Figs. 23 and 24 at respective lines 1052 and 1100. Looking now in
detail to Fig. 25, x-Channel ~2n signals, as welt as y-Channel Qn are submitted from
each respective row Rnd column readout through input conduits, represented
generaUy at 1110 and 1112, to a Coincidence Network and x, y Pair Code Generatorfunction 1114. NetwDrk lll4 provides a read-in function which checks the inputs at
1110 and 1112 corresponding to a given gamms ray interaction within a given detector
component for the coincidence oI their time derivative signals. When the
30 coincidence of such a data pair is received, thè spatial position represented by such
signals is assured and a corresponding x, y Pair Code is generated and presentedthrough appropriate conduits represented generally by the transfer conduit at 1116.
This pair code output is inserted into F.I.F.O. Asynchronous Memory and Line
Decoder function 1118 by virtue of a clocked signal or pulse provied through line 1120.
F.LF.O. Asynchronous Memory 1118 corres~onds with the same function provided at
516 in Fig. 19. As before, the F.I.F.O. (first-in, first-out) memory is conventionally
--63--
~:1 3~3~
formed incorporating generally in(lependent input and output stages or networks. It
serves witllin the system as a de-randornizer which receives and collects or records
the x, y Pair Codes from circuit 1114 and, following a four-to-three line decoding
thereof3 submits signals to multiplexers 510-514 (Fig. 19) providing for the selective
acceptance of the signals addressed thereto. Note in the latter regard, that thesignal labeling for the instant embodiment remains the same as that shown in Fig.
19. These coded instructions to the noted multiplexers are represented in Fig. 25 by
the broad conduit arrows appropriately labeled and respectively identified at 1122-
1126 .
lû Returning to the Coincidence ~etwork and Pair Code Generator function 1114,
in the absence of a noted signal coincidence identifying a proper spatial pair code,
appropriate signal return channel alignment will be provided through a multi~channel
conduit represented generally at 1128 to a Reset Control function 1130. Operating in
similar fashion at reset drive function 520 in Fig. 19, control 1130 responds to a non-
coincidence condition to provide for the resetting of appropriate ones of the row or
colunn readout networks as described in connection with Figs. 23 and 24. In thisregard, note that multiple row reset outputs are represented generally by arrow 1132,
while the corresponding column or y-Channel output signals are directed through a
conduit arrangement represented generally at 1134. As noted above, the signals
20 labeled at the latter two outputs are representative of multi-row and multi-column
interconnections. Reset signal transmitting control over Reset Co~3trol 1130 is
derived from Sequential Control block 1140 as through the conduit represented
generally at 1142. Control 1140 additionally provides the functions described inconnection with block 526 in ~ig. 19, i.e clocking the information codes from
F.I.F.O. Aæynchronous Memory and Line Decoder 1118 by outputs submitted thereto
through line 1144; controlling the Sample and Hold Amplifi~rs 544, 548 and 552 to
receive information from Multiplexers 510 and 514 and assert appropriate delays
suited for the proper operation of Two Channel Analyzer 562, so as to assure that a
signal of proper energy level criteria is processed. Further, function 1140 activateæ
30 Reset Control ~130 to generated end-of-cycle resetting as well as short cycleresetting performance occasioned with the failure of a given signal to pass the
window criteria of analyzer 562. The outputs of Control 1140 to the noted Sampleand Hold Amplifiers are represented generally by the broad arrow at 1146, while the
corresponding input thereto from the processing system providing for the noted
resetting or recycling features is represented by the broad arrow designated 1148.
--64--
With the noted replacement of processing control 522' for that earlier
represented at 522 in Fig. 19, the sytem operates essentially in the same manner,
i.e., multiple channel analysis being carried out over the several energy levels which
may be provided by components of the general type described at 562; sample and
hold functions are carried out, as described at 544-552; and a dividing function is
provided to normalize the signals with respect to their corresponding energy levels
by divider functions as at 584 and 586 in Fig. 19.
Since certain changes may be made in the system and apparatus without
parting from the scope of the invention herein involved, it is intended that all10 matter contained in the above description or shown in the accompanying drawings
shall be interpreted as illustrative and not in a limiting sense.
--~5-
