Physiology & Behavior, Vol. 35, pp. 395-403. Copyright© PergamonPress Ltd., 1985. Printedin the U.S.A. 0031-9384/85$3.00 + .00
A Microcomputer-Based Method for
Physiologically Interpretable
Measurement of the Rewarding Efficacy of
Brain Stimulation
KENNETH A. CAMPBELL, 2 GARY EVANS 3 AND C. R. GALLISTEL
Department of Psychology, University of Pennsylvania, Philadelphia, PA 19104
Received 13 August 1984
CAMPBELL, K. A., G. EVANS AND C. R. GALLISTEL. A microcomputer-based method for physiologically interpre-
table measurement of the rewarding efficacy of brain stimulation. PHYSIOL BEHAV 35(3) 395-403, 1985.-
Determination of the function relating rate of pressing to the number of pulses in a train of fixed duration (the rate-
frequency function) yields a physiologically interpretable measure of changes in the rewarding efficacy of the stimulation,
because the number of action potentials in the reward-relevant first stage axons is directly proportional to the number of
pulses in the train. We describe a system, based on a low cost microcomputer, which permits determination of 16-data-
point rate-frequency functions in 4--6 animals simultaneously in less than 10 minutes. We give an empirical and theoretical
justification for using the curve-shift measurement procedure in drug and lesion work, where the experimental treatments
must be presumed to have substantial effects on performance factors.
Microcomputer Brain-stimulation-reward Quantitative methods Performance-vs.-reward
IDENTIFICATION of the neural substrate for brain stimu-
lation reward requires a behavioral measurement technique
that satisfies two requirements: it should distinguish changes
in rewarding efficacy from changes in other factors that af-
fect the performance of the rewarded behavior; and it should
yield quantitative estimates of the effect on rewarding effi-
cacy that can be related to physiological measures. The re-
warding effect of stimulation is the level of reward that it
produces. By contrast, the rewarding efficacy of stimulation
is its capacity to produce a criterial level of reward. Meas-
ures of changes in this capacity reveal quantitative charac-
teristics of the neural substrate that carries the rewarding
signal from the electrode to the site where the rewarding
effect is realized [10]. Behavioral measurements of elec-
trophysiological, anatomical and pharmacological properties
provide quantitative data about the substrate, against which
to compare the results from electrophysiological, anatomical
and neurochemicat measurements.
It has long been recognized that changes in the rate of
responding do not provide a suitable measure of changes in
the rewarding efficacy of the stimulation [13,23]. Firstly,
changes in the rate may occur in response to changes in
performance factors---the motor, perceptual and motiva-
tional processes that determine how an underlying rewarding
effect maps into an observed performance. Secondly, even if
a change in the rate could be ascribed to a change in reward-
ing efficacy, the magnitude of the behavioral change would
not measure the magnitude of the underlying physiological
change, that is, the magnitude of the change in the neural
signal that generates the rewarding effect. One could not
conclude that a drug that reduced the rate of responding by a
factor of two reduced the rewarding efficacy of stimulation
by a factor of two.
Most of the behavioral methods that have been suggested
for distinguishing non-specific performance effects from ef-
fects on reward in drug and lesion studies yield at best a qual-
itative answer to the question whether the drug or lesion has
altered the rewarding efficacy of the stimulation [6, 11, 21].
These methods do not attempt to measure the magnitude of
the underlying physiological effect. Measurements of the
self-stimulation current threshold [4] come closer to a quan-
titative approach. However, any measure that looks at only a
single fixed level of performance may confound changes in
performance factors with changes in rewarding efficacy. An
overall depression or elevation of performance will produce
an artifactual depression or elevation in the amount of stimu-
1This research was supported by NSF Grant BNS 82 11972 to CRG, to whom requests for reprints, copies of the disks, circuit diagrams, and
other documentation should be addressed. The authors are grateful to Gwen Freyd and Elysa Braunstein for the data used in Figs. l and 3.
2Now at Department of Physiology; Bowman Gray School of Medicine; Winston-Salem, NC 27103; supported as a postdoctoral fellow on
NIMH Training Grant T32 MH15092, during the period of this work.
3Now at Simple Solutions; 4108 Pavilion Place; Durham, NC 27707.
395