EFFECTS OF D-AMPHETAMINE IN A TEMPORAL DISCRIMINATION PROCEDURE: SELECTIVE CHANGES IN TIMING OR RATE DEPENDENCY?

JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR
2002, 78, 195–214
NUMBER 2 (SEPTEMBER)
EFFECTS OF D-AMPHETAMINE IN
A TEMPORAL DISCRIMINATION PROCEDURE:
SELECTIVE CHANGES IN TIMING OR
RATE DEPENDENCY?
AMY L. ODUM, LORI M. LIEVING, AND DAVID W. SCHAAL
UNIVERSIT Y OF NEW HAMPSHIRE AND
WEST VIRGINIA UNIVERSIT Y
Two experiments evaluated rate dependency and a neuropharmacological model of timing as ex-
planations of the effects of amphetamine on behavior under discriminative control by time. Four
pigeons pecked keys during 60-trial sessions. On each trial, the houselight was lit for a particular
duration (5 to 30 s), and then the key was lit for 30 s. In Experiment 1, the key could be lit either
green or blue. If the key was lit green and the sample was 30 s, or if the key was lit blue and the
sample was 5 s, pecks produced food on a variable-interval 20-s schedule. The rate of key pecking
increased as a function of sample duration when the key was green and decreased as a function of
sample duration when the key was blue. Acute d-amphetamine (0.1 to 3.0 mg/kg) decreased higher
rates of key pecking and increased lower rates of key pecking as predicted by rate dependency, but
did not shift the timing functions leftward (toward overestimation) as predicted by the neurophar-
macological model. These results were replicated in Experiment 2, in which the key was lit only one
color during sessions, indicating that the effects were not likely due to disruption of discriminative
control by key color. These results are thus consistent with rate dependency but not with the pre-
dictions of the neuropharmacological model.
Key words: timing, rate dependency, temporal discrimination, amphetamine, key peck, pigeons
In a neuropharmacological model of time
erence memory, and if the ratio of difference
perception, Meck (1996) proposed that do-
between the two is lower than a threshold val-
pamine and acetylcholine provide the basis
ue, a response occurs.
for the operation of major components of
Meck (1996) suggested that dopamine lev-
scalar expectancy theory (SET; Gibbon, 1977;
els determine the clock process, primarily
Gibbon & Church, 1984; Gibbon, Church, &
through in?uencing pacemaker rate (Hinton
Meck, 1984). Brie?y, SET is an information-
& Meck, 1997). The more dopamine present,
processing model of timing that posits clock,
the faster the clock ticks. An increase in do-
memory, and decision stages. In the clock
pamine levels is predicted to produce an im-
stage, pulses from a pacemaker are gated
mediate leftward shift in the psychophysical
through a switch into an accumulator. The
timing function (i.e., the animal would over-
number of pulses in the accumulator repre-
estimate the passage of time and respond pre-
sents the amount of time that has passed
maturely). A decrease in dopamine levels, on
since the switch closed and timing began.
the other hand, is predicted to produce an
The value from the accumulator can be trans-
immediate rightward shift in the psychophys-
ferred to reference memory in the memory
ical timing function (i.e., the animal would
stage when an important event, such as the
underestimate the passage of time and re-
delivery of food for a deprived animal, oc-
spond later than usual).
curs. In the decision stage, the value from the
Meck (1996) suggested that levels of acetyl-
accumulator is compared to the value in ref-
choline determine temporal memory, primar-
ily through in?uencing the translation con-
We thank Jennifer Johnson for assistance in conduct-
stant that modi?es the value transferred from
ing the experiment and Timothy Shahan for helpful
the clock stage to the memory stage (Hinton
comments during the conduct of the experiment and on
& Meck, 1997). The more acetylcholine pres-
previous drafts of this manuscript.
Portions of this research were presented at the 26th
ent, the shorter the remembered duration of
annual meeting of the Association for Behavior Analysis,
events. An increase in acetylcholine levels is
Washington, DC, May 2000.
predicted to produce a gradual leftward shift
Address correspondence to Amy Odum, Department
in the psychophysical timing function (i.e., the
of Psychology, Conant Hall—10 Library Way, University
of New Hampshire, Durham, New Hampshire 03824 (e-
animal would gradually come to overestimate
mail: aodum@hopper.unh.edu).
the passage of time and respond premature-
195

---

196
AMY L. ODUM et al.
ly). A decrease in acetylcholine levels, on the
1968), rats (e.g., Ksir & Nelson, 1977; Mc-
other hand, is predicted to produce a gradual
Auley & Leslie, 1986), mice (e.g., Glowa,
rightward shift in the psychophysical timing
1986; McKim, 1980), monkeys (e.g., Goethe
function (i.e., the animal would gradually
& Isaac, 1977; Herling, Downs, & Woods,
come to underestimate the passage of time
1979), and humans (e.g., Stitzer, 1984) main-
and respond later than usual).
tained by FI schedules.
The present experiments examined the ef-
This increase in responding early in the in-
fects of d-amphetamine on the behavior of
terval has been interpreted to re?ect overes-
pigeons in a temporal discrimination proce-
timation of the passage of time (e.g., Killeen,
dure. In addition to other neurochemical ef-
1991; Maricq et al., 1981; McAuley & Leslie,
fects, amphetamine increases dopamine lev-
1986; Meck, 1996). Within behavioral phar-
els by stimulating the release of dopamine
macology, however, these same types of
from presynaptic terminals and blocking the
changes in the temporal patterning of re-
reuptake of dopamine from the synapse (Sei-
sponding within ?xed intervals have long
den, Sabol, & Ricaurte, 1993). Furthermore,
been considered an example of rate depen-
although there are some anatomical differ-
dency (e.g., Dews & Wenger, 1977; Kelleher
ences in the avian and mammalian forebrain,
& Morse, 1968; McKearney & Barrett, 1978;
the function and organization of the dopa-
see Odum & Schaal, 2000, for discussion).
mine system is similar in birds (including pi-
Rate dependency in its most basic form is the
geons) and mammals (including rats; Dur-
empirical generalization that the effects of a
stewitz, Kro
¨ner, & Gu
¨ ntu
¨ rku
¨ n, 1999; Vischer,
drug on behavior depend on the rate of the
Cue´nod, & Henke, 1982). Amphetamine has
behavior in the absence of the drug (Dews,
other effects besides increasing dopamine
1981). For behavior maintained by FI sched-
(Seiden et al., 1993), and more speci?c com-
ules, drugs from a variety of pharmacological
pounds for altering dopamine levels exist. A
classes increase low rates of behavior early in
substantial amount of the evidence that Meck
the interval, but particularly with higher dos-
(1996) offers for the role of dopamine in tim-
es, decrease higher rates of behavior later in
ing, however, comes from experiments inves-
the interval.
tigating the effects of amphetamine (e.g.,
The dif?culty in differentiating between a
Maricq & Church, 1983; Maricq, Roberts, &
temporal overestimation account and a rate
Church, 1981; Meck, 1983). The current ex-
dependency account of the effects of am-
periments therefore focus on the effects of
phetamine on behavior maintained by FI
amphetamine, because this drug has been
schedules is that the two accounts make the
used to provide support for the neurophar-
same prediction: Low early rates should in-
macological model (Meck, 1996) and be-
crease in both cases. The present experi-
cause amphetamine has a long history in the
ments attempted to evaluate these two ac-
investigation of the effects of drugs on timing
counts in a novel way by arranging a situation
(e.g., Schuster & Zimmerman, 1961; Sidman,
in which they make divergent predictions.
1955).
Based on a procedure described by Reynolds
In addition to other forms of evidence to
and Catania (1962), pigeons were presented
support the role of dopamine in timing,
with a sample duration and then were given
Meck (1996) cited experiments showing that
the opportunity to peck a response key for 30
amphetamine increases low response rates
s. A range of sample durations was presented
maintained in the initial portion of ?xed-in-
across trials. If the key was lit one color, pecks
terval (FI) schedules of reinforcement. In an
were intermittently reinforced following the
FI schedule, the ?rst response after a ?xed
shortest sample duration. If the key was lit
time produces access to a reinforcer (Ferster
another color, however, pecks were rein-
& Skinner, 1957, pp. 133–134). Response
forced only following the longest sample du-
rates typically increase across the interval
ration.
from low or zero rates early in the interval to
Figure 1 shows how this procedure differ-
higher steady rates near the end of the inter-
entiates between the predictions of overesti-
val. Amphetamine produces robust increases
mation and rate dependency accounts of the
early in the interval in response rates of pi-
effects of amphetamine on behavior. Under
geons (e.g., Katz & Barrett, 1979; McMillan,
control conditions (top panel), response

---

CHANGES IN TIMING OR RATE DEPENDENCY?
197
decrease as a function of sample duration in
the component in which pecks can produce
food following short sample durations. The
corresponding time on the x axis at which the
two functions cross is the point of subjective
equality (PSE): the sample duration that is
perceptually midway between the longest and
shortest sample and therefore that at which
the response rate is the same in the two com-
ponents. The rest of the panels show various
potential disruptions of timing. If a drug were
to produce overestimation of the passage of
time (second panel), as amphetamine is sug-
gested to do (Meck, 1996), then the func-
tions would be shifted horizontally, to the left
on the x axis, and the PSE would be reduced.
Underestimation (third panel) would shift
the functions horizontally, to the right on the
x axis, and the PSE would be increased. If the
effects of a drug were rate dependent (fourth
panel), however, the PSE should remain the
same, but lower rates should increase and
higher rates should decrease.
The present experiments evaluated the
role of dopamine in timing as posited by
Meck (1996). In Experiment 1, d-amphet-
amine was administered to pigeons respond-
ing on the procedure described above. If in-
creases in dopamine increase clock speed,
then the timing functions should shift to the
left (horizontally). If the effects are largely
rate dependent, however, the functions
should not shift to the left, but should instead
?atten. Experiment 2 evaluated whether the
results of Experiment 1 could re?ect a dis-
ruption in color discrimination rather than
changes in timing by arranging the long and
short components separately across condi-
tions.
EXPERIMENT 1
Fig. 1.
Explanatory ?gure with hypothetical data
showing response rate as a function of sample duration
METHOD
in the current procedure for control performance and
for several possible changes that could be produced by
Subjects
drugs (see text). Circles show data from trials in which
responses could be reinforced following the longest sam-
The subjects were 4 adult male White Car-
ple; squares show data from trials in which responses
neau pigeons with previous acute exposure to
could be reinforced following the shortest sample.
cocaine (Schaal, Miller, & Odum, 1995) and
morphine (Odum, Haworth, & Schaal, 1998;
Odum & Schaal, 1999) while responding un-
rates should increase as a function of sample
der a multiple ?xed-ratio FI schedule or a
duration in the component in which pecks
multiple FI clocked FI schedule of food pre-
can produce food following long sample du-
sentation. The pigeons were maintained at
rations, much like during an FI, but should
80% (
10 g) of their free-feeding weights

---

198
AMY L. ODUM et al.
through postsession feedings of mixed grain
center key was lit either blue or green for 30
as necessary. Pigeons received no drugs for
s. If the houselight duration was 5 s and the
40 to 46.5 weeks prior to tests of d-amphet-
key was lit blue, key pecks were reinforced
amine in the current experiment. When not
during this period according to a variable-in-
in experimental sessions, pigeons were indi-
terval (VI) 20-s schedule (composed of 20 in-
vidually housed in a temperature-controlled
tervals generated using the constant-proba-
colony under a 12:12 hr light/dark cycle and
bility method of Catania & Reynolds, 1968,
were allowed free access to water and diges-
and the BASIC program of Perone; see Lattal,
tive grit. Sessions were conducted during the
1991). If the houselight duration was 5 s and
light portion of the cycle.
the key was lit green, key pecks were not re-
inforced. If the houselight duration was 30 s,
Apparatus
however, pecks when the key was green were
Four custom-made experimental cham-
reinforced according to the same VI 20-s
bers, constructed of wood with aluminum
schedule (i.e., there was only one VI schedule
front panels, were used. The internal dimen-
that operated under both circumstances), but
sions were 33 cm across the front panel, 31
pecks when the key was blue were not rein-
cm from the front panel to the back wall, and
forced. The VI schedule operated only follow-
37.5 cm from the ?oor to the ceiling. Three
ing the 5-s sample duration when the key was
plastic response keys (2.1 cm diameter) on
blue and following the 30-s sample duration
the front panel were mounted 26 cm from
when the key was green. The time during
the ?oor. The center key could be lit from
hopper presentations was not included in the
behind with green or blue light and required
30-s period of access to the lit key. When 30
a force of approximately 0.19 N to record a
s elapsed, the center keylight was extin-
response. The side keys were dark, and pecks
guished and the houselight was lit to begin
to these keys had no programmed conse-
the next trial (i.e., there was no intertrial in-
quences. A lamp (28 V, 1.1 W) 7 cm above
terval).
the center key served as the houselight. A
To avoid extinguishing key pecking during
rectangular aperture 16 cm below the center
the initial exposure to the procedure, during
key provided access to a solenoid-operated
the ?rst session the only durations present
food hopper ?lled with mixed grain. White
were those following which responses could
noise and chamber ventilation fans masked
be reinforced (5 s for blue and 30 s for
extraneous sounds. Contingencies were pro-
green). Intermediate sample durations (10,
grammed and data collected by microcom-
15, 20, and 25 s) were added one at a time
puters located in an adjacent room with Med
in ascending order for blue and descending
Associates interfacing and software.
order for green across the next four sessions.
Procedure
Finally, during the sixth session, the ?nal tim-
Experimental sessions were conducted dai-
ing procedure was reached in which each
ly at approximately the same time. Due to the
sample duration (5, 10, 15, 20, 25, and 30 s)
pigeons’ previous history, no key pecking or
occurred prior to both key colors. Pecks to
hopper pretraining was necessary. Reinforce-
the center key after the presentation of the
ment was 3-s access to the food hopper. Dur-
intermediate sample durations did not pro-
ing hopper presentations, the aperture was lit
duce food in either type of trial. Each sample
with white light, and the houselight and key-
duration was presented 10 times, ?ve times
light were extinguished. Ten minutes after
preceding the blue center key and ?ve times
the pigeons were placed in the darkened
preceding the green center key, for a total of
chamber, the session began with the lighting
60 trials. The order of trial types was random-
of the houselight.
ly chosen each session. In summary, pecks to
The procedure was essentially a multiple-
the blue key could be reinforced following
schedule version of the timing procedure of
the shortest houselight presentation (5 s; the
Reynolds and Catania (1962). Two types of
short component) and pecks to the green key
trials alternated randomly. Trials began when
could be reinforced following the longest
the houselight was lit for either 5 or 30 s.
houselight presentation (30 s; the long com-
Then the houselight was turned off and the
ponent).

---

CHANGES IN TIMING OR RATE DEPENDENCY?
199
Drug Administration
was lit blue and food was available after the
5-s sample).
Pigeons experienced 180 to 198 sessions of
The top row of Figure 3 shows temporal
the ?nal timing procedure prior to adminis-
discrimination functions during control ses-
tration of d-amphetamine. Drug testing be-
sions for each pigeon. Mean rates of key
gan for individual pigeons when response
pecking decreased as a function of sample
rates in both components were stable for
duration during the short component. Dur-
each sample duration as judged by visual in-
ing the long component, mean rates of key
spection (i.e., showed no increasing or de-
pecking increased as a function of sample du-
creasing trends or extreme variability over
ration. For each pigeon, the two functions
the ?nal 15 to 20 sessions). A range of doses
crossed between the geometric mean (12.5 s)
of d-amphetamine (0.1, 0.3, 1.0, and 3.0 mg/
and the arithmetic mean (17.5 s) of the two
kg) and its vehicle (0.9% saline) were admin-
extreme sample durations. The lower rows in
istered in a mixed order for each pigeon. d-
Figure 3 show the effects of increasing doses
Amphetamine (Sigma) was dissolved in saline
of d-amphetamine. The smallest dose of am-
and administered in a volume of 1.0 ml/kg
phetamine (0.1 mg/kg; second row) had no
of the body weight at 80% of free-feeding
large or systematic effects on the rates of key
weight. Amphetamine and vehicle were ad-
pecking as a function of sample duration. As
ministered via intramuscular injections into
in control sessions, rates of key pecking in-
the breast immediately before the pigeons
creased as a function of sample duration in
were placed in the experimental chambers.
the long component and decreased as a func-
Tests were separated by at least three consec-
tion of sample duration in the short compo-
utive baseline sessions not preceded by injec-
nent. Increasing doses of amphetamine in-
tions. The session immediately preceding a d-
creased lower rates of key pecking and
amphetamine or vehicle test session was
decreased higher rates of key pecking for
designated a control session. Dose–effect
each pigeon (i.e., the functions relating rate
curves were determined completely before
of key pecking to sample duration ?attened).
any dose was repeated and four dose–effect
The point at which the long and short func-
cur ves were completed. Pigeons were
tions crossed, however, did not change sys-
weighed prior to and after experimental ses-
tematically across pigeons. The PSE did not
sions, and drug tests were not conducted if
appear to change for P78, shifted to the left
initial weights were not within 10 g of the ap-
for P40, and shifted to the right for P10 and
propriate weight for the pigeon. This event
P53.
rarely occurred.
Figure 4 allows assessment of the rate-de-
R
pendent effects of amphetamine on key peck-
ESULTS
ing by expressing the data from Figure 3 in
Figure 2 shows the overall rate of key peck-
another format. Mean rates for each sample
ing (i.e., rates averaged across those obtained
duration following d-amphetamine adminis-
following all sample durations) as a function
tration were divided by control rates for those
of dose of d-amphetamine for each pigeon
sample durations for each dose of d-amphet-
separately for both components. Saline had
amine for each component separately for
little systematic effect on the overall rate of
each pigeon. The resulting number was then
key pecking. With increasing doses of am-
multiplied by 100 and plotted on logarithmic
phetamine, rates of key pecking generally in-
axes as a function of the mean control rate
creased moderately and then decreased.
during the corresponding sample duration.
Rates for P40, however, did not decrease at
Rates were largely unaffected by the admin-
the largest dose in either component, and
istration of 0.1 mg/kg d-amphetamine (top
rates did not decrease at the largest dose for
row). The next highest doses (0.3 and 1.0
P10 and P53 in the short component. Across
mg/kg) elevated low rates somewhat, but usu-
pigeons, mean rates of key pecking were not
ally had little effect on higher rates. At the
systematically different between the long
highest dose (3.0 mg/kg), higher rates were
component (in which the key was lit green
decreased for all birds, with lower rates re-
and food was available after the 30-s sample)
maining elevated for 3 of 4 birds. In other
and the short component (in which the key
words, the effect of d-amphetamine on key

---

200
AMY L. ODUM et al.
Fig. 2.
Effects of injections of d-amphetamine on the mean rate of key pecking per session in the short component
(left panels) and the long component (right panels) for each pigeon. Unconnected points show means for all control
(C) and saline (S) sessions. Lines connect points showing mean rates for four determinations of the effects of each
dose. Vertical bars indicate one standard deviation around means. In some cases the variability around a point is
encompassed by the point. The horizontal line originating from the y axis shows the control mean for comparison
purposes. For clarity, y axes are scaled for individual pigeons.
pecking depended on the usual rate of key
creased the rate of key pecking that usually
pecking. In general, the effects were similar
occurred at a low rate, whereas higher doses
for rates in the long component and rates in
of d-amphetamine decreased key pecking
the short component. Amphetamine in-
that usually occurred at a higher rate.

---

CHANGES IN TIMING OR RATE DEPENDENCY?
201
Fig. 3.
Mean rates of key pecking as a function of sample duration in the long component (green key; triangles)
and short component (blue key; circles) during 16 control sessions (top row) and following administration of four
determinations of the effects of increasing doses of amphetamine (descending rows) for each pigeon (columns).
Vertical bars indicate one standard deviation around means. For reference, dashed lines indicate one standard de-
viation above and below means from control sessions preceding injections (control means shown top row). For clarity,
y axes are scaled for individual pigeons.
Figure 5 compares mean rates of key peck-
trials with food and trials without food. Rates
ing per session on S
trials only (i.e., trials on
during trials without food tended to decrease
which food was potentially available; 5 s for the
at lower doses of amphetamine than rates dur-
short component and 30 s for the long com-
ing trials with food. In some instances there
ponent) for trials during which food was de-
was substantial variability across replications,
livered and for trials during which food was
however, and the effect was in some cases
not delivered. There were fewer trials without
small or inconsistent across pigeons or doses.
food per session, and data from these trials
were in some cases more variable than data
DISCUSSION
from trials with food. For both the long and
In the absence of d-amphetamine, behavior
short components, during control and saline
showed evidence of discriminative control by
sessions rates of key pecking were similar for
time. The rate of key pecking as a function

---

202
AMY L. ODUM et al.
Fig. 4.
Response rates during each sample duration following administration of d-amphetamine for each of 4
pigeons (columns) at four doses of amphetamine (rows). The x axis shows mean control rate during each sample
duration, and the y axis shows percentage of that control rate during the same sample duration following amphet-
amine administration. The horizontal line indicates 100% of control rate (i.e., no change relative to control). Points
above the line represent rate increases and points below the line represent rate decreases, relative to control rates.
Circles depict rates during the short component, and triangles depict rates during the long component. Linear
regression lines were ?t by the method of least squares. For clarity, x axes are scaled for individual pigeons. See text
for details of calculations.
of sample duration in control sessions (Fig-
obtained by Reynolds and Catania (1962)
ure 3, top row) showed evidence of temporal
when the key was lit only one color and food
discrimination: Rates increased as a function
was available following either the longest or
of sample duration when food was available
shortest sample duration across conditions.
following the longest sample, and rates de-
Furthermore, in the present experiment the
creased as a function of sample duration
functions relating rates of key pecking to sam-
when food was available following the short-
ple duration for the long and short compo-
est sample. These results are similar to those
nents crossed between the geometric and

---

CHANGES IN TIMING OR RATE DEPENDENCY?
203
Fig. 5.
Effects of injections of d-amphetamine on mean rates of key pecking for trials with food (?lled circles)
and trials without food (open circles) following the 5-s sample (short; left column) or 30-s sample (long; right column)
for each pigeon. For clarity, y axes are scaled for individual pigeons and points are offset slightly. Other details are
as in Figure 2.

---

204
AMY L. ODUM et al.
arithmetic means of the temporal endpoints.
served at higher doses on trials with food.
This result indicates that the PSE was between
Thus, the ?attening of the functions relating
the arithmetic and geometric mean, as found
rates of key pecking to sample duration can-
previously in different temporal discrimina-
not be due to the presence of food on S
tion procedures (e.g., Stubbs, 1968; see also
trials, because the timing functions would
Killeen, Fetterman, & Bizo, 1997).
tend to be even ?atter if trials with food were
Two lines of evidence suggest that the dis-
excluded.
crimination was in fact based on the sample
The present results are therefore inconsis-
duration, not on the presence or absence of
tent with the neuropharmacological model of
food. First, under control conditions, the
timing proposed by Meck (1996). One pos-
functions relating rate of key pecking to sam-
sibility, however, is that the results do not re-
ple duration were graded: Rates increased
?ect a disruption in temporal discrimination,
gradually as a function of sample duration in
but rather a disruption in stimulus control by
the green component and decreased gradu-
the color of the response key. Temporal con-
ally as a function of sample duration in the
trol could have been maintained, or changed
blue component (Figure 3, top row). If the
in a way consistent with the neuropharma-
discrimination were based on the presence or
cological model, but once the sample ended,
absence of food, one would expect a step
the bird could then not discriminate whether
function: low equal rates after all S
stimuli
the key was lit blue or green. A failure to dis-
(i.e., stimuli following which food was never
criminate the color of the response key could
available) and high rates after S
stimuli.
plausibly produce a general ?attening of the
Furthermore, under control conditions, rates
function relating the rate of key pecking to
of key pecking on S
trials did not differ sys-
sample duration. Experiment 2 was conduct-
tematically on trials with food and trials with-
ed to examine this interpretation.
out food (Figure 5).
The effects of d-amphetamine on overall
EXPERIMENT 2
response rates (Figure 2) were similar to
those obtained previously for behavior main-
To investigate the possibility that the results
tained by interval schedules of positive rein-
obtained in Experiment 1 re?ected a break-
forcement: At lower doses, amphetamine had
down in stimulus control by the color of the
no effect or increased the overall rate of key
response key, rather than purely changes in
pecking, but at higher doses, amphetamine
temporal discrimination, this experiment ar-
generally decreased the overall rate of key
ranged the components across conditions
pecking (see van Haaren, 1993). Amphet-
rather than within sessions. In one condition,
amine did not, however, produce the shifts in
the response key was lit red and pecks could
the timing functions that are predicted by the
be reinforced only after the shortest sample
neuropharmacological model of timing
duration (5 s). In another condition, the re-
(Meck, 1996). Amphetamine increases levels
sponse key was lit white and pecks could be
of dopamine and should therefore shift the
reinforced only after the longest sample du-
timing functions to the left, indicating over-
ration (30 s). This procedure therefore did
estimation of elapsed time. In the present ex-
not require the color of the response key to
periment, however, amphetamine more com-
be discriminated within sessions, and any
monly increased lower rates of key pecking
changes in response patterning could be
and decreased higher rates of key pecking
more plausibly attributed to disruption of
(Figure 4), thus ?attening the timing func-
temporal control. Furthermore, because the
tions rather than shifting them horizontally
lowest dose of d-amphetamine tested in Ex-
(Figure 3). The results thus indicate a general
periment 1 had no effect on behavior, the
decrease in the control of behavior by time,
dose range was shifted to slightly higher dos-
rather than overestimation of the passage of
es.
time. Examining the effects of d-amphet-
METHOD
amine on rates of key pecking during S
tri-
als only (Figure 5), to the extent that rates of
Subjects
key pecking differed on trials with food and
The pigeons from Experiment 1 were
trials without food, rates tended to be pre-
transferred to the University of New Hamp-

---