Age and Sex Differences in Behavioral Flexibility, Sensitivity to Reward Value, and Risky Decision-Making Assignment | Get Paper Help

ill upload to files. You have to read the article and just follow the APA format fileAge and Sex Differences in Behavioral Flexibility, Sensitivity to Reward
Value, and Risky Decision-Making
Sara R. Westbrook, Emily R. Hankosky, Megan R. Dwyer, and Joshua M. Gulley
University of Illinois, Urbana-Champaign
Compared with adults, adolescent behavior is often characterized by reduced behavioral flexibility,
increased sensitivity to reward, and increased likelihood to take risks. These traits, which have been
hypothesized to confer heightened vulnerability to psychopathologies such as substance use disorders
(SUDs), have been the focus of studies in laboratory animal models that seek to understand their neural
underpinnings. However, rodent studies to date have typically used only males and have adopted standard
methodological practices (e.g., weight loss inducing food restriction) that are likely to have a disparate
impact on adolescents compared with adults. Here, we used adolescent and adult Sprague–Dawley rats
of both sexes to study instrumental behavior tasks that assess behavioral flexibility (strategy shifting and
reversal learning; Experiment 1), sensitivity to reward value (outcome devaluation; Experiment 2), and
risky decision making (probability discounting; Experiment 3). In Experiment 1, we found that adolescents
were faster to acquire reversal learning than adults but there were no differences in strategy shifting.
In Experiments 2 and 3, adolescents and adults were equally sensitive to changes in reward value and
exhibited similar reductions in preference for a large reward when reinforcement probability was
decreased. However, adolescents responded more efficiently and earned reinforcers at a higher rate than
their same-sex, adult counterparts. Together, these findings provide only limited support for the existence
of an “adolescent-typical” phenotype in Sprague–Dawley rats and instead suggest that age differences in
the expression of these behaviors may depend on conditions such as pubertal status and motivational
state.
Keywords: adolescence, risk-taking, behavioral flexibility, reward sensitivity
Adolescence, which is the transitional period between childhood
and adulthood that is characterized by physical, socioemotional,
hormonal, and behavioral changes, is a time when psychopathologies
such as substance use disorders (SUDs) often emerge (Chen,
Storr, & Anthony, 2009). Prominent theories of adolescent behavior
propose that delayed neurobehavioral development leads to a
relative imbalance between the early maturing reward system and
the late-maturing cognitive control system, which underlies an
individual’s self-regulation abilities (Casey, Jones, & Somerville,
2011; Shulman, Harden, Chein, & Steinberg, 2015; Steinberg,
2008; Steinberg, 2010). This imbalance is hypothesized to contribute
to the adolescent-typical behaviors, such as increased risk
taking, that are associated with risk for developing SUDs (Jupp &
Dalley, 2014; Mitchell et al., 2014). Behavioral flexibility, one
component of cognitive control, is the ability to inhibit inappropriate
behaviors to promote more adaptive ones in response to
environmental changes (Brown & Tait, 2014). In humans, behavioral
flexibility tends to increase from late childhood to young
adulthood with girls performing better than boys at younger adolescent
ages and boys increasing performance across adolescence
(Kalkut, Han, Lansing, Holdnack, & Delis, 2009). Age is negatively
associated with risk taking both in laboratory tasks of risky
choice (Cauffman et al., 2010; Eshel, Nelson, Blair, Pine, & Ernst,
2007; Gardner & Steinberg, 2005) and in self-report measures
(Gardner & Steinberg, 2005). In addition, males tend to make more
risky choices than females (van Leijenhorst, Westenberg, & Crone,
2008). Sensation- or reward-seeking peaks in adolescence and
males generally exhibit higher levels than females (Shulman et al.,
2015). Moreover, pubertal status and age have been shown to be
dissociable in humans, as pubertal status positively correlates with
adolescent sensation-seeking even when controlling for age (Martin
et al., 2002). Collectively, these results are in line with a
hypersensitive reward system and protracted maturation of the
cognitive control system in adolescents and highlight the influence
of sex in these behaviors.
Studies that utilize rodent models of adolescence, which have
potential advantages for investigating underlying neural mechanisms
of age differences in behavior, have typically focused on
responses to psychoactive drugs (Anker, Baron, Zlebnik, & Carroll,
2012; Anker & Carroll, 2010; Hankosky, Westbrook, Haake,
Marinelli, & Gulley, 2017; Kantak, Goodrich, & Uribe, 2007;
This article was published Online First February 26, 2018.
Sara R. Westbrook, Emily R. Hankosky, and Megan R. Dwyer, Department
of Psychology, University of Illinois, Urbana-Champaign; Joshua M.
Gulley, Department of Psychology, and Neuroscience Program, University
of Illinois, Urbana-Champaign.
The authors thank Laura Cortes, Kristen Hughes, Shawn Kurian, Sarah
Rahman, and Rebecca Sandoval for excellent technical assistance. This
work was supported by the National Institutes of Health Grant DA 029815
to Joshua M. Gulley.
Correspondence concerning this article should be addressed to Joshua
M. Gulley, Department of Psychology and Neuroscience Program,
University of Illinois, Urbana-Champaign, 731 Psychology Building
MC-716, 603 East Daniel Street, Champaign IL 61820. E-mail:
jgulley@illinois.edu
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Behavioral Neuroscience © 2018 American Psychological Association
2018, Vol. 132, No. 2, 75–87 0735-7044/18/$12.00 http://dx.doi.org/10.1037/bne0000235
75
Shahbazi, Moffett, Williams, & Frantz, 2008; Shram, Funk, Li, &
Le, 2008). These studies using drug reinforcers can be challenging
to interpret, however, because it is difficult to separate the contribution
of adolescent developmental changes in general reward and
cognitive control processes from those induced by the drugs themselves.
Notably, adult rats that self-administer psychostimulants
have been shown to have deficits in reward processing (Green,
Dykstra, & Carelli, 2015), impaired behavioral flexibility (Cox et
al., 2016), and increased risk taking (Mitchell et al., 2014). An
alternative approach that avoids these potentially confounding
drug effects on behavior is to substitute nondrug reinforcers in the
analysis of potential age effects.
Rodent studies of age differences in behavioral flexibility that
use positive (food) or negative (escape from water) reinforcement
have reported mixed results (Newman & McGaughy, 2011; Simon,
Gregory, Wood, & Moghaddam, 2013; Willing, Drzewiecki,
Cuenod, Cortes, & Juraska, 2016), with factors such as pubertal
onset potentially contributing to disparate findings (Willing et al.,
2016). Studies of age differences in reward processing also highlight
the importance of the peripubertal period in males, as there is
a peripubertal peak in motivation for (Friemel, Spanagel, & Schneider,
2010) and consumption of palatable reward (Friemel et al.,
2010; Marshall, Liu, Murphy, Maidment, & Ostlund, 2017). In the
only study to concurrently investigate age and sex differences in a
Pavlovian approach task, our lab found that females and adolescents
were less sensitive to reward devaluation (Hammerslag &
Gulley, 2014). In adults, sex differences have been reported where
females were insensitive to devaluation of a food reinforcer
(Quinn, Hitchcott, Umeda, Arnold, & Taylor, 2007). One explanation
for reduced response suppression following reward devaluation
is that adolescents and females maintain a greater value
representation of the outcome due to their relative hypersensitivity
to reward. However, our previous study only examined age and
sex differences when the reward was devalued. To date, a single
rodent study in males reported age differences in risky choice, with
adolescents exhibiting less sensitivity to reward probability compared
with adults (Zoratto, Laviola, & Adriani, 2013). Sex differences
in risky choice have only been examined in adults, with
males generally displaying more risky behavior than females (for
review see Orsini & Setlow, 2017). Thus, although there appears
to be emerging evidence of age and sex differences in reward
sensitivity, behavioral flexibility, and risk taking, the current literature
is mostly limited to either male or adult subjects, respectively,
which precludes analysis of interacting effects of age and
sex in these behavioral domains.
The goal of the present study was to address these gaps by using
adolescent and adult Sprague–Dawley rats of both sexes to concurrently
examine age and sex differences in behavioral flexibility
using an operant strategy-shifting task (Experiment 1), reward
value sensitivity in an instrumental outcome devaluation task
(Experiment 2), and risk-based decision making in a probability
discounting task (Experiment 3). In line with leading theories of
adolescent vulnerability that propose adolescents are hypersensitive
to reward and exhibit poor cognitive control (Casey et al.,
2011; Shulman et al., 2015; Steinberg, 2010), we predicted that
adolescents would display less behavioral flexibility, enhanced
sensitivity to reward value, and increased risky choices relative to
adults. In addition, consistent with previous rodent studies using
nondrug reinforcers (Hammerslag & Gulley, 2014; Orsini & Setlow,
2017; Quinn et al., 2007), we predicted that these differences
would be more pronounced in males relative to females.
Method
Subjects
Subjects were 147 Sprague–Dawley rats (69 male, 78 female)
born in-house from breeders originally obtained from Envigo
(Indianapolis, IN). Rats were weaned on postnatal day (P) 22 and
housed two to three per cage in a temperature-controlled room on
a 12-hr reversed light/dark cycle (lights off at 9:00 a.m.). Rats were
weighed daily beginning on P25 (Experiments 1 and 3) or P33
(Experiment 2). A subset of the rats (38 male, 43 female) were
checked daily for markers of pubertal onset, which are vaginal
opening in females and preputial separation in males, (Castellano
et al., 2011; Korenbrot, Huhtaniemi, & Weiner, 1977). These
checks began on P30 and continued until all rats reached puberty.
In this sample, average pubertal onset was estimated to be at 34.9
0.2 days for females (range: 32–39 days) and 43.6 0.2 days
for males (range: 41–47 days).
Water was available in the home cage ad libitum throughout the
study. Food was available ad libitum until the start of the behavioral
training when it was limited to 20 hr/day. Specifically, food
was removed from the home cage 2 hr prior to behavioral sessions
and was returned 0.5 hr–2 hr after their conclusion. Rats maintained
on this regimen exhibit sufficient motivation during instrumental
behavior sessions, but maintain body weights that are not
significantly different from those in age-matched, free-fed controls
during the periadolescent period of rapid growth (Figure 1A). All
experimental procedures were approved by the Institutional Animal
Care and Use Committee at the University of Illinois, Urbana-
Champaign, and followed the Guide for the Care and Use of
Laboratory Animals (National Research Council, 2011).
Apparatus
Rats were trained and tested in standard operant chambers
(Coulbourn Instruments, Whitehall, PA) that were housed inside
sound-attenuating cubicles. The cubicles were equipped with fans
that provided ventilation and masked extraneous noise. One wall
of each operant chamber was equipped with a centrally located
food trough outfitted on either side with a response module (retractable
levers in Experiment 1, Experiment 2; recessed nosepoke
ports in Experiment 3) that were equidistant (87 mm) from
the trough. White cue lights were located above each response
module and a white houselight was located near the chamber
ceiling on the wall opposite the food trough. Sessions were recorded
and analyzed using Graphic State software (Coulbourn
Instruments, Allentown, PA).
Experiment 1: Strategy Shifting
The timeline for training and testing for the 23 male and 25
female rats used in Experiment 1 is shown in Figure 1B. Training
began with one session of “magazine training,” during which 40
pellets (45 mg; Bio-Serv F0021 Dustless Precision Pellets) were
delivered individually into the food trough on a random time 100-s
schedule. Subsequently, rats were trained during four 30-min
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
76 WESTBROOK, HANKOSKY, DWYER, AND GULLEY
sessions (one session/day) to respond on both levers on a fixed
ratio 1 (FR1) schedule of reinforcement. The order in which the
levers were presented alternated each day. Rats that did not acquire
lever pressing were manually shaped using successive approximations.
Following lever press training, rats were trained during
seven sessions to respond within 10 s of lever extension into the
chamber. Each rat’s side bias was then established in a single
session as described by Floresco, Block, and Tse (2008).
Daily strategy-shifting sessions consisted of 120 trials separated
by a 20-s intertrial interval (ITI). During each trial, one of the two
cue lights located above the levers was illuminated and both levers
were extended into the chamber. Individual cue lights were presented
pseudorandomly across trials, such that no cue light could
be presented on more than two consecutive trials. Rats were first
trained to use a visual strategy by reinforcing them for pressing the
lever that had a cue light illuminated above it, regardless of the
lever’s spatial location (i.e., left or right side of pellet delivery
trough). An incorrect choice resulted in retraction of the levers,
10-s illumination of the houselight, and a return to the ITI. Trials
continued until rats achieved eight consecutive correct choices and
had completed at least 30 trials.
On the day following acquisition of this performance criterion,
or after a maximum of five visual strategy training sessions, rats
began sessions where they could complete a within-session shift to
an egocentric response strategy. This within-session shift, which
occurred following eight consecutive correct choices using the
visual strategy, required rats to respond according to lever position
(e.g., the left lever), regardless of cue light location, for reinforcement
to occur. The location of the reinforced lever for the response
strategy was determined individually for each rat such that it was
the lever opposite of the rat’s side bias observed during initial
training. Trials continued until rats achieved a criterion of eight
consecutive correct choices or a maximum of 120 trials. Rats
received up to a maximum of three sessions to complete the
within-session shift. Rats were shifted from the visual strategy to
the response strategy as shifting in this direction has been shown
to be medial prefrontal cortex (PFC) dependent, while the opposite
shifting direction was not (Floresco et al., 2008).
Following acquisition of the response strategy or the maximum
of three sessions, rats received two daily response strategy sessions
consisting of 75 trials each. One day following these sessions, the
response strategy was reversed using a within-session design. The
session started with the original response strategy reinforced until
rats had eight consecutive correct lever choices. Subsequently, the
rule was reversed such that they were reinforced for pressing on
the previously nonreinforced lever (e.g., the right lever). As before,
trials continued until rats achieved a criterion of eight consecutive
correct choices. For all stages of the task, trials where rats failed to
respond on a lever within 10 s of trial onset were terminated,
scored as an omission, and were followed by a 10-s illumination of
the houselight before the ITI commenced. Omitted trials were
repeated until rats made a response within the required time.
Experiment 2: Outcome Devaluation
The timeline for training and testing for the 28 male and 32
female rats used in Experiment 2 is shown in Figure 1B. Rats
began training with two daily 60-min sessions of preexposure to
two food reinforcers: 45-mg food pellets (Bio-Serv F0021 Dustless
Precision Pellets) and 66.7% sweetened condensed milk (SCM;
Meadow Gold). In these sessions, which were intended to reduce
neophobia, rats were individually placed in cages and given free
access to the food reinforcers for 30 min each. The following day,
rats began training in the operant chambers with a single session of
“magazine training” that lasted for 30 trials. For these sessions, the
houselight was off, levers were retracted, and the food trough was
illuminated for 5 s when a reinforcer was delivered on a random
time 60 s (RT-60). Reinforcer type (food pellet or SCM) was
counterbalanced across rats and a single food pellet or 0.04 ml
dipper cup of SCM was delivered for each trial. Training continued
with two sessions of autoshaping that each involved up to 25
reinforcer deliveries as a result of lever pressing (30 min duration).
For each trial, the houselight and a cue light above a single
extended lever was illuminated and the position of the lever (left or
right of the food trough) was counterbalanced across rats. When a
response occurred (FR1) or a time interval elapsed (RT-60), the
lever was retracted, the houselight was extinguished, and the
assigned reinforcer was delivered along with illumination of
the food trough for 5 s. During two additional training sessions on
subsequent days, reinforcers were only delivered following a lever
Figure 1. Summary of body weight gain during the periadolescent period
and the experimental timeline for the study (A). Shown are data from rats
in the adolescent groups that were allowed access to food for 20 hr/day
during training and testing and those from the adult groups that had not yet
begun the restricted access procedure and thus had food available 24
hr/day. The data points for rats in the 24-hr access groups are plotted with
a nudge of 0.3 x-axis units so they are not obscured by the data points from
the restricted groups. Separate two-way ANOVAs revealed significant
main effects of postnatal day (males: F(30, 1615) 466, p .0001;
females: F(30, 1871) 305, p .0001). In males, there was a significant
age by group interaction, F(30, 1615) 1.88, p .0028, reflecting that
weights were only significantly different between groups on P49 ( p
.05). (B) Experimental timeline showing the training and testing ages for
each experiment. Mean age of pubertal onset for each sex in the subset of
rats that underwent puberty checks is indicated by their respective symbols.
See the online article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
AGE AND SEX DIFFERENCES IN INSTRUMENTAL BEHAVIOR 77
press (FR1). Rats that made fewer than 10 responses during these
sessions were manually shaped for 30 min.
Following this initial training, rats received sessions with reinforcement
on a random/variable interval 30-s (RI-30) schedule that
ended after 50 outcomes were earned, 20 min elapsed without a
lever press, or 90 total min elapsed. Three RI-30 sessions preceded
Extinction Test Set 1 and three additional RI-30 sessions occurred
prior to Extinction Test Set 2. A single test set consisted of two
extinction test days (one of each value condition) separated by one
retraining RI-30 session. On extinction test days, rats received a
60-min prefeeding session in which they were given ad libitum
access to either the instrumental outcome (devalued condition) or
the alternative reinforcer (valued condition). Immediately following
prefeeding, rats were placed in the operant chambers for a
10-min test under extinction conditions in which lever presses
were recorded but had no consequence. Following the extinction
test, rats were returned to their prefeeding cages for a 15-min
posttest consumption where they had simultaneous access to both
reinforcers. The order of outcome value condition for the test days
was counterbalanced across rats.
Experiment 3: Probability Discounting
The timeline for training and testing for the 18 male and 21
female rats used in Experiment 3 is shown in Figure 1B. During
the two days prior to the first training session, rats were given 25
food pellets (45 mg; Bio-serv F0021 Dustless Precision Pellets) in
their homecage on two successive days. Training began with one
session of magazine training during which the nose-poke ports
were covered and the houselight extinguished. A food pellet was
delivered and the magazine light illuminated for 5 s on an RT-30
reinforcement schedule for a total of 30 deliveries. On the following
day, the nose-poke ports were uncovered and assignment of the
ports to a small (one food pellet) or large (three food pellets)
reward was counterbalanced across rats. Illumination of a green
light inside one port signaled that the port was active and rats were
trained to nose-poke on an FR1 schedule for a total of 25 reinforcements/
session or 30 min had elapsed. Only one port was
illuminated per session and the active port was alternated for four
daily sessions. If the rat failed to nose-poke within 10 s following
illumination of the active port, the trial was scored as an omission,
and the houselight was illuminated to begin the 20-s ITI to repeat
the trial.
Test sessions consisted of 12 forced choice trials followed by 12
free choice trials. During the forced choice trials, one active port
was illuminated randomly each trial for a total of six presentations
of each port. During the free choice trials, both ports were active
and illuminated. Omitted trials were repeated following the ITI.
Responses in the small, certain reward port were reinforced with
delivery of one pellet at 100% probability. Responses in the large,
potentially risky reward port were reinforced with delivery of three
pellets at 100%, 66.7%, 33.3%, or 16.7% probability that descended
across sessions. In 23 rats (10 male, 13 female), four
sessions at 100% were followed by three sessions each at 66.7%,
33.3%, and 16.7% probabilities. In the remaining 16 rats (eight
male, eight female), the procedure was modified slightly so that
they received 100% sessions with only six forced choice trials until
each rat in the group showed a preference (higher than 50%
choice) for the large, risky port before they were allowed to
proceed to the three sessions at each of the other probabilities. This
change, which resulted in the first cohort of rats having four
sessions at 100% probability and the second having seven to eight
sessions, was implemented due to concerns about satiation from
the large number of pellets rats received during the 100% probability
forced choice trials. However, subsequent analysis of the test
session data with age, sex, probability, and cohort as factors reveal
there was no main effect of cohort, F(1, 31) 3.54, p .0695,
and no interactions with cohort (all F’s 2.16; all p’s 0.15).
Thus, the data from these two cohorts were grouped for all subsequent
analyses.
Data Analysis
Dependent measures used to assess cognitive flexibility (Experiment
1) included the number of trials to criterion, errors to
criterion, and omissions for each strategy (visual, response, and
response reversal). Dependent measures for each strategy were
analyzed with two-way ANOVA (Age Sex). The relationship
between trials to criterion performance on each strategy and the
subsequent strategies was assessed using linear regression analyses.
For outcome devaluation (Experiment 2), dependent measures
were rate of lever pressing and reinforcement rate during acquisition
and the RI-30 sessions, consumption during the feeding sessions
before and after extinction tests, and lever presses during
extinction tests. Paired t tests were conducted on lever presses
during the extinction test sets to reveal any statistically significant
differences in lever presses between the value conditions. Significance
of the obtained p values was evaluated after first adjusting
for multiple comparisons using false discovery rate (less than 5%;
Benjamini & Hochberg, 1995). A mixed factorial ANOVA was
conducted on lever presses during the test sets (Age Sex Test
Set Value Condition). Mixed factorial ANOVAs were conducted
on lever pressing rate (lever presses/min) and rate of
reinforcer earned (g/kg/min) during the acquisition period (Age
Sex Session) and RI-30 sessions (Age Sex Session).
Prefeeding consumption (g/kg) on extinction test days was analyzed
using mixed factorial ANOVA (age and sex as between
subject factors; test set and reinforcer as within subject factors).
For the posttest consumption (g/kg) data, separate factorial
ANOVAs for pellets and SCM were conducted (Age Sex
Test Set Value Condition as Factors).
The dependent measure used to assess risk-based decision making
(Experiment 3) was the percent choice of the large reward
nose-poke port during the free choice trials on the last session at
each probability of large reinforcement. A three-way mixed
ANOVA was conducted with age, sex, and probability (within
subjects) as factors. Reinforcement rate adjusted for body weight
(g/kg/min) and omissions during the free choice trials were also
analyzed for the final session at each probability using a three-way
mixed ANOVA. To assess reward discrimination, two-way
ANOVA (Age Sex) was conducted on the percent large reward
choice and reinforcement rate (g/kg/min) in free choice trials on
the final session of 100% probability of large reinforcement. For
all analyses, Tukey post hoc comparisons were conducted when
appropriate.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
78 WESTBROOK, HANKOSKY, DWYER, AND GULLEY
Results
Experiment 1: Strategy Shifting
A total of 23 male and 25 female rats began Experiment 1, but
one rat was excluded from the final analyses due to a technical
problem (n 1 adult female). Thus, final group sizes were 12 each
for adolescent females and adolescent males, 12 adult females, and
11 adult males. Shown in Figure 2 are the total number of trials
required to meet the performance criterion for each strategy. For
the initial strategy on which rats were trained (visual), a total of 12
rats failed to achieve the acquisition criterion of eight consecutive
correct (n 2 adolescent males, n 2 adolescent females, n 5
adult males, n 3 adult females), resulting in a broad performance
range of as few as 111 trials to the maximum of 960 trials (which
was the maximal number of visual strategy trials a rat could
receive). Despite this range, two-way ANOVA of the group data
for both trials and errors to criterion for the visual strategy revealed
no significant main effects or interactions (Figure 2A; errors to
criterion not shown). Linear regression analyses demonstrated that
trials to criterion performance on the visual strategy did not significantly
predict subsequent trials to criterion performance for the
egocentric response strategy (Adj R2 0.0077, p .25) or the
response reversal (Adj R2 0.0191, p .17). Thus, we present
performance on these strategies including the data from all rats,
whether or not they met acquisition criterion for the visual strategy.
Following the visual strategy, rats were required to shift to an
egocentric response strategy. Compared with the initial strategy,
Figure 2. Trials to criterion (TTC) and omissions during acquisition of the (A, B) visual strategy, (C, D)
response strategy, and (E, F) response reversal in the strategy-shifting task. Rats (n 11–12/group) met the
acquisition criterion when they achieved eight consecutive correct responses. Those that failed to meet the
criterion for the visual strategy were assigned the total number of trials they received, which was 960. p .05
versus adults. See the online article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
AGE AND SEX DIFFERENCES IN INSTRUMENTAL BEHAVIOR 79
the response strategy was learned by all rats in considerably fewer
trials that ranged from eight to 114 trials (Figure 2C). Still, there
were no group differences in response strategy performance, as
two-way ANOVA of trials and errors to criterion revealed no
significant main effects or interactions. Subsequently, when the
contingency for the response strategy was reversed such that rats
had to make an alternate response (e.g., press the left lever instead
of the right), there were age differences in performance (Figure
2E). Two-way ANOVA of trials to criterion revealed a significant
main effect of age, F(1, 43) 4.21, p .046, with adolescents
requiring significantly fewer trials to reach criterion than adults.
Linear regression analysis showed that the number of trials required
to reach criterion for the response strategy did not significantly
predict the trials to criterion for the subsequent response
reversal (Adj R2 0.0451, p .08). In addition, all rats met the
acquisition criterion for the reversal. Analysis of total errors to
criterion for the response reversal revealed no significant main
effects or interactions.
Total omissions during each strategy also are presented in
Figure 2. There were no significant differences in omissions during
the visual (Figure 2B) and response reversal strategies (Figure 2F).
However, adolescents had more omissions during the egocentric
response strategy compared to adults (Figure 2D), age: F(1, 6)
6.64, p .042. Importantly, omissions did not explain differences
in performance as there were no differences in trials to criterion for
the response strategy.
Experiment 2: Outcome Devaluation
A total of 28 male and 32 female rats began Experiment 2.
However, 10 rats were excluded from the final analyses due to
their failure to learn to lever press (n 3 adolescent females, n
2 adolescent males), failure to meet the criterion of at least one
lever press during the test sessions (n 1 adolescent female, n
3 adult male), or a technical problem (n 1 adult female). Thus,
the final group sizes were 13 adolescent females, 12 adolescent
males, 14 adult females, and 11 adult males.
Lever pressing rate was assessed across two autoshaping and
two lever press training sessions during the initial lever press
acquisition period (Figure 3A). Reinforcer type (food pellet or
SCM) was not used as a factor because mixed ANOVA revealed
there was no significant main effect of reinforcer type and it did
not interact with age, sex, or session. We observed significant main
effects of age, F(1, 46) 23.65, p .0001; and session, F(3,
46) 100.48, p .0001, as well as a significant Age Session
interaction, F(3, 46) 11.87, p .0001. Post hoc analysis
revealed no age difference in lever pressing rate during autoshaping
(Sessions 1 and 2), but adults had a higher lever pressing rate
than adolescents during both lever press training sessions, Session
3 (p .0387) and Session 4 (p .0001). For both adolescents and
adults, response rate increased across these initial training sessions
(adolescents: from Sessions 2–4, all p’s 0.03; adults: each
session, all p’s 0.01). Because body size differs considerably
between adolescents and adults, and this may influence both capacity
and motivation to make lever press responses that lead to
reinforcement, we also analyzed responding during these RI-30
sessions with a body weight-adjusted rate of reinforcement measure
(g of reinforcer/kg/min; Figure 3C). This analysis indicated a
main effect of session, F(3, 46) 36.10, p .0001, such that rats
increase their reinforcement rate across all sessions (all p’s
0.01). In addition, adult females had higher reinforcement rates
than adult males (p .027), while adolescent reinforcement rates
were similar in both sexes (Age Sex: F(1, 46) 4.23, p .045).
Lever pressing rate during RI-30 sessions is shown in Figure 3B.
Across these sessions, all rats (regardless of age and sex) increased
their rate of lever pressing (session: F(7, 46) 13.70, p .0001)
and adolescents pressed the lever at a slower rate than adults (age:
F(1, 46) 6.60, p .014). Mixed factorial ANOVA also revealed
a significant Age Session interaction, F(7, 46) 2.64, p .022.
Specifically, adolescents pressed the lever at a slower rate than
adults during the sessions prior to the first test set (Sessions 1–3)
and Session 5. However, this age effect diminished or no longer
existed in the RI-30 sessions that followed the first devaluation test
set (Sessions 6–8). In these sessions, adolescents responded at
higher rates similar to those in adults. Analysis of reinforcement
rate during RI-30 sessions revealed that adolescents (age: F(1,
46) 110.64, p .0001) and females (sex: F(1, 46) 63.42, p
.0001) had a higher rate of reinforcement compared to adults and
males, respectively. Mixed factorial ANOVA also revealed a significant
main effect of session, F(7, 46) 4.74, p .001 (Figure
3D, significance not shown). Specifically, rats earned the reinforcers
at the lowest rate in Session 1 (p .011 vs. Session 2) but this
rate remained from Session 2 onward.
During the prefeeding phase of the devaluation tests, rats were
given unlimited access to consume one of the reinforcers for 60
min. A mixed factorial ANOVA of the consumption data during
the prefeeding sessions across Test Sets 1 and 2 revealed there was
no main effect of test set so these data were collapsed. As shown
in Figure 4, all rats consumed more SCM than pellets during the
prefeeding sessions (main effect of reinforcer: F(1, 46) 489, p
.0001). In addition, females consumed more reinforcer than males
(main effect of sex: F(1, 46) 18.9, p .0001), but the significant
Sex Age interaction, F(1, 46) 10.6, p .0022, and the
subsequent post hoc tests revealed this effect was driven primarily
by the consumption differences in adult females compared to adult
males.
Immediately following the prefeeding phase, rats were given a
brief extinction test and lever presses during these sessions are
shown in Figure 5. Rats displayed a significant reduction in lever
presses when the reinforcer was devalued compared with valued
(value condition: F(1, 184) 62.4, p .0001). Paired t tests
revealed that all groups exhibited a significant reduction in lever
presses when the reinforcer was devalued compared with valued
(p’s 0.05), with the exception of adolescent females in the
second test set. In addition, mixed factorial ANOVA revealed
several significant interactions (Age Sex: F(1, 184) 6.09, p
.015; Age Test Set: F(1, 184) 4.04, p .049; Age Sex
Test Set: F(1, 184) 3.91, p .049). These interactions resulted
because adult males had fewer lever presses than adolescent males
during the second test set collapsed across value condition, but this
effect was not present in females (Figure 5B).
Following the extinction tests, rats were given simultaneous
access to both reinforcers for 15 min in order to evaluate the
efficacy of the devaluation procedure (see Figure 6). As we observed
for the prefeeding phase, there was no significant main
effect of test set so these data were collapsed for subsequent
analyses. Importantly, each group consumed fewer pellets when
pellets were devalued through prefeeding than when the value of
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
80 WESTBROOK, HANKOSKY, DWYER, AND GULLEY
pellets remained intact through prefeeding of the alternate reinforcer
(Figure 6A; value condition: F(1, 46) 375, p .0001).
This pattern, which was also evident in SCM consumption (Figure
6B; value condition: F(1, 46) 284, p .0001), suggests that
prefeeding was a successful manipulation to change the value of
the reinforcers. Despite all groups being sensitive to a change in
value of both rewards, adult females consumed more pellets than
adult males when the value of pellets was intact (Age Sex
Value Condition: F(1, 46) 5.29, p .026; Sex Value
Condition: F(1, 46) 4.33, p .043), while adolescents consumed
more SCM than adults when SCM was devalued (Age
Value Condition: F(1, 46) 6.11, p .0172; Age: F(1, 46)
36.54, p .0001). Females also consumed more SCM than males,
regardless of value condition (Sex: F(1, 46) 9.45, p .0035).
Experiment 3: Risk-Based Decision Making
A total of 18 male and 21 female rats began and subsequently
completed Experiment 3, yielding group sizes of 11 adolescent
females, eight adolescent males, 10 adult females, and 10 adult
males. Percent choice of the large reward port and reinforcement
rate adjusted for body weight are shown in Figure 7. Rats reduced
their choice of the large reward as the probability of reinforcement
decreased (Figure 7A; probability: F(3, 105) 8.46, p .0001).
On the last day when the probability of large reinforcement was
100%, two-way ANOVA of percent large reward choice revealed
that adults tended to make the large reward choice more than
adolescents (age: F(1, 35) 4.17, p .0487), but this group
difference was modest and appeared to be largely driven by adult
males (Figure 7B).
Reinforcement rate adjusted for body weight was examined for
the free choice trials in the final session at each probability (Figure
7C). Adolescents had higher reinforcement rates than adults at all
probabilities (age: F(1, 35) 36.41, p .0001). Reinforcement
rates of rats was significantly reduced when the probability of
large reinforcement decreased to 16.7% (probability: F(3, 105)
7.67, p .0001). In addition, females earned more reinforcer than
Figure 3. Responding during the instrumental outcome devaluation task (n 11–14/group). Rates of lever
pressing (lever presses/min) are shown during the acquisition period (A) and the RI-30 sessions (B) with breaks
in the graph corresponding to extinction testing days. p .05 and p .001 versus adolescents, collapsed
across sex. (C) Rates of reinforcement adjusted for body weight (g/kg/min) are shown for the acquisition period.
# p .05 adult females versus adult males. During the RI-30 sessions (D), there were main effects of age (p
.0001) and sex (p .0001) for rates of reinforcement. See the online article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
AGE AND SEX DIFFERENCES IN INSTRUMENTAL BEHAVIOR 81
males (sex: F(1, 35) 68.68, p .0001). On the final day of the
100% probability of large reinforcement, two-way ANOVA of
reinforcement rate revealed that adolescents (Figure 7D; age: F(1,
35) 14.26, p .001) and females (sex: F(1, 35) 20.08, p
.0001) had higher reinforcement rates than adults and males,
respectively.
Analysis of omissions during free choice trials on the final day
of each probability revealed no significant main effects or interactions
(data not shown). Further analyses were performed to
address concerns of potential group differences in satiation within
a session. Mixed factorial ANOVAs (Age Sex Trial Port)
were conducted to examine omissions and latency to nose-poke
upon illumination of the ports during the free choice trials of the
final session where the large reward had a 100% probability of
reinforcement. These analyses revealed no significant main effects
or interactions, with the exception of a significant main effect of
port, F(1, 28) 4.58, p .0411, on the latency to nose-poke upon
port illumination, reflecting that rats took longer to nose-poke into
the large reward port compared to the small reward port.
Discussion
In the present study, we tested rats under minimal, weight
maintaining food restriction conditions in multiple instrumental
behavior tasks that allowed investigation of potential age- and
sex-dependent differences in behavioral flexibility, reward value
sensitivity, and risk-based decision making. In the strategy-shifting
task, we found that adolescent rats, regardless of sex, exhibited
faster acquisition of reversal learning, suggesting adolescents have
greater behavioral flexibility than adults. In both the outcome
devaluation and risk-based decision making tasks, adolescents of
both sexes responded in a more efficient manner than their samesex
adult counterparts, earning reinforcers at a significantly higher
rate. However, adolescent and adult rats of both sexes were equally
sensitive to devaluation, suggesting no age difference in updating
outcome value representations. The only sex difference in this task
was in adults, with adult females consuming more reward than
adult males during the prefeeding session. In addition, we found
that adolescents and adults, regardless of sex, made similarly risky
choices in the probability discounting task. Overall, our results are
only partially consistent with our a priori hypotheses and the
predictions of leading theories of adolescent vulnerability.
In the strategy-shifting task, we found that adolescent rats displayed
enhanced behavioral flexibility compared to adult rats,
regardless of sex. This effect was specific to reversal learning, as
adolescents acquired the extradimensional shift from visual to
response strategies just as rapidly as adults. It is important to note,
however, that rats in the present study required considerably more
trials to acquire the visual strategy than those in previous studies
(Cox et al., 2016; Floresco et al., 2008), which could be due to
strain differences in visual acuity (Kumar, Talpos, & Steckler,
2015). Like Kumar, Talpos, and Steckler’s (2015) study, some of
our Sprague–Dawley rats failed to acquire the initial visual discrimination.
Notably, we found that performance on the visual
strategy did not significantly predict subsequent performance in
the shift to response strategy. The direction of shift from the visual
Figure 4. Reward consumption during the prefeeding phase of the outcome
devaluation experiment. Consumption data for pellets and sweetened
condensed milk (SCM) are adjusted for body weight and averaged across
the two test sets. p .0001 versus adult females, within reinforcer type.
See the online article for the color version of this figure.
Figure 5. Total number of lever presses during the 10-min extinction sessions are shown for Test Set 1 (A) and
Test Set 2 (B). Each test set consisted of two extinction sessions immediately following the prefeeding phase.
The extinction test days were separate by one RI-30 session. p .05, p .01 versus devalued within group
in paired t test; @ p .05 versus adolescent males collapsed across value condition. See the online article for
the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
82 WESTBROOK, HANKOSKY, DWYER, AND GULLEY
strategy to the egocentric response strategy has been shown to be
medial PFC-dependent due to the cost of switching from an
easier to a more difficult strategy (Floresco et al., 2008).
Because the visual strategy in the present study seemed to be
rather difficult for our rats to acquire, the strategy shift to
response in our task may not have been sensitive to medial PFC
function and this may explain the similar performance between
adolescents and adults.
Figure 6. Post extinction test consumption (g/kg) of pellets (A) and SCM (B) by value condition. During the
prefeeding phase, rats were given the reinforcer they expected in the test session (devalued condition, which was
pellets for A and SCM for B) or the alternate reinforcer (valued condition, which was SCM for A and pellets
for B). p .0001 versus devalued within group; @ p .05 versus adult females with pellets valued; ### p
.0001 versus adults with SCM devalued. See the online article for the color version of this figure.
Figure 7. Responses in the large reward nose-poke port during free-choice trials of the risky decision-making task.
Data (n 8–11/group) are presented as the percent of large reward choices on (A) the last session when each
probability of reinforcement was tested and (B) during the last session when the probability of reinforcement was fixed
at 100%. Rate of reinforcement adjusted for body weight (g/kg/min) is shown for (C) the last session at each
probability of reinforcement for the large reward choice with main effects of age (p .0001) and sex (p .0001),
and (D) during the final session at 100% probability. p .05, p .001 versus 100.0%, collapsed across age and
sex; p.05, p.05 versus 66.7% collapsed across age and sex; ### p.001 versus females collapsed across
age; @ p .05 versus adolescents. See the online article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
AGE AND SEX DIFFERENCES IN INSTRUMENTAL BEHAVIOR 83
The lack of age differences in the extradimensional shift is
consistent with a previous study (Newman & McGaughy, 2011)
when adolescents were tested at similar ages (P50 and P53) as our
experiment (P44 –50). In addition, our finding that adolescents
exhibit greater behavioral flexibility in reversal learning compared
to adults is consistent with some previous findings (Simon et al.,
2013), but not others (Newman & McGaughy, 2011). Differences
in study methodology, such as stimulus complexity, rat strain, food
restriction, and adolescent pubertal status may explain the disparity.
Newman and McGaughy (2011) employed the intradimensional/
extradimensional attentional set-shifting task in which rats
were required to dig in pots for food reinforcers. Stimulus complexity
in this task may have influenced their finding of adolescents
needing more trials to acquire a compound discrimination
(Newman & McGaughy, 2011); adolescents may be more susceptible
to distraction by irrelevant stimuli. The pots in this study
varied in multiple dimensions making them a much more complex
stimulus set compared to a tone (Simon et al., 2013) or light cue
(the present study). This complexity may have contributed to the
earlier study’s finding that adolescent rats displayed less cognitive
flexibility compared with adults. Differences in stimulus complexity
may also lead to differences in switch cost due to task difficulty.
Regarding rat strain, Newman and McGaughy (2011) used
Long Evans rats to show adolescents were less flexible than adults,
whereas we and Simon et al. (2013) found adolescent Sprague–
Dawley rats were more flexible than adults. Adolescents and adults
are differentially sensitive to motivational factors, such as food
restriction, as adolescents exhibited more perseverative responding
during extinction compared with adults when food deprived (Sturman,
Mandell, & Moghaddam, 2010). Newman and McGaughy
(2011) food restricted rats to 90% free-feeding weight of agedmatched
controls, whereas those in the present study were fed ad
libitum 20 hr/day and experienced no weight loss compared with
rats fed ad libitum 24 hr/day. The greater food restriction employed
by Newman and McGaughy (2011) relative to that in the
present study may have contributed to their finding of adolescent
behavioral inflexibility by increasing perseverative responding.
Lastly, regarding the influence of pubertal stage, it was recently
reported that male and female Long Evans rats who reached
puberty within days of training in a Morris water maze task had
significantly better performance on the reversal learning stage of
the task compared to prepubertal rats (Willing et al., 2016). Thus,
despite similar testing ages for the adult groups in ours and the
previous studies (Newman & McGaughy, 2011; Simon et al.,
2013), the age and presumed pubertal status of the adolescent
groups may contribute to inconsistent findings. Future studies are
required to pinpoint the precise role that these various methodological
factors play in laboratory rodent assessments of age and
sex differences in behavioral flexibility.
Many adolescent changes in neural regions associated with
behavioral flexibility seem to be most prominent during the peripubertal
period. Previous work has shown that behavioral flexibility
is sensitive to changes in PFC function (Dalton, Wang, Phillips,
& Floresco, 2016; Floresco et al., 2008; Ghods-Sharifi, Haluk, &
Floresco, 2008), with different subregions controlling distinct
components of flexible behavior. Specifically, inactivation of the
medial PFC impaired extradimensional set-shifts, while sparing
(Floresco et al., 2008) or even enhancing reversal learning (Dalton
et al., 2016). In contrast, inactivation of the orbitofrontal cortex
(OFC) impaired reversal learning, while sparing set-shifts (Ghods-
Sharifi et al., 2008). Substantial neuroanatomical changes, including
pruning of neurons (Markham, Morris, & Juraska, 2007; Willing
& Juraska, 2015) and synapses (Drzewiecki, Willing, &
Juraska, 2016), have been reported to occur at pubertal onset in the
medial PFC in rats of both sexes. These neuroanatomical changes
may underlie adult-like performance in set-shifts in the postpubertal
adolescent rats in the present study and the previous literature
discussed above (Newman & McGaughy, 2011). However, it
remains unclear whether there are peri-pubertal neuroanatomical
changes in the OFC, and if they play a role in age-dependent
differences in reversal learning.
The current study and other recent findings (Hankosky et al.,
2017; Marshall et al., 2017) suggest that reward sensitivity in
adolescent rodents is likely dependent on multiple factors, including
motivational state and pubertal status. During the acquisition
period for lever pressing in Experiment 2, adolescents had a lower
response rate compared to adults. This effect persisted during the
RI-30 sessions leading up to the first test set. Although the response
rate was lower in adolescents, an age difference in rate of
reinforcement emerged during the RI-30 sessions. Adolescents did
not differ from adults in their rate of reinforcement during the
acquisition period, but displayed a higher rate of reinforcement
compared with adults throughout the RI-30 sessions. These results
may suggest that adolescent rats are more behaviorally efficient
under the RI-30 schedule of reinforcement as they respond less
frequently during this schedule but still earned reinforcers at a
higher rate than their adult counterparts. Adolescents’ reduced
response rate during random interval schedules is consistent with
a previous study (Andrzejewski et al., 2011), although those authors
did not report analysis of reinforcement rate normalized to
body weight. The lower response rate in adolescents may be due to
more physical effort required to lever press compared with adults.
However, our findings from Experiment 3 where the required
response was a nose-poke rather than a lever press argue against
this as we found a similar age difference in rate of reinforcement.
Thus, even when the effort cost of the response is reduced (a
nose-poke vs. a lever press), age differences in reinforcement rate
persist. Together, our results highlight the importance of analyzing
age differences in reinforcement in a manner that takes age differences
in body size into account.
Despite evidence for a peri/postpubertal peak in reward consumption
(Friemel et al., 2010; Marshall et al., 2017), motivation
(Friemel et al., 2010), and cocaine self-administration in males
(Wong, Ford, Pagels, McCutcheon, & Marinelli, 2013), several
studies measuring instrumental responding for nondrug reinforcers
report either comparable responding (Kim, Simon, Wood, &
Moghaddam, 2016; Naneix, Marchand, Di Scala, Pape, & Coutureau,
2012; Sturman & Moghaddam, 2011) or decreased responding
in adolescents compared to adults (present study; Andrzejewski
et al., 2011; Hankosky et al., 2017; Sturman et al., 2010).
One factor that could contribute to the apparent discrepancies
between reward consumption and instrumental responding is
whether the measure accounts for differences in body weight.
Studies of reward consumption and drug self-administration adjust
for differences in body weight, while measures of instrumental
responding for nondrug reinforcers typically do not. As such,
studies of sex differences in nondrug reward processing have
found that females consume more reward per body weight when
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
84 WESTBROOK, HANKOSKY, DWYER, AND GULLEY
given ad libitum access (present study; Hammerslag & Gulley,
2014; Marshall et al., 2017), but seem to respond for reinforcer
similarly or more than males (present study; Hankosky et al.,
2017) depending on whether the measure is normalized to body
weight. In the present study, as the adolescents increase in body
size across RI-30 sessions, they increase their responding in order
to maintain their elevated reinforcement rate (adjusted for body
weight) relative to adults. Studies examining reward consumption
normalized to body weight have consistently found that adolescent
males consume more reward than their adult counterparts (Friemel
et al., 2010) regardless of caloric content (Marshall et al., 2017).
While these data suggest that adolescent reward hypersensitivity
does not depend on age differences in caloric need, other studies
that account for body weight but use noncaloric reinforcers do not
clearly support this interpretation (Hankosky et al., 2017; Shram et
al., 2008). Future studies are needed to understand how age differences
in caloric need due to body size may influence instrumental
responding.
Despite age differences in response and reinforcement rate during
RI-30 sessions leading up to the first test set, we found no
group differences in sensitivity to reward devaluation during the
first test set (when adolescent rats were P42–P44), which coincides
with mean pubertal onset in our sample of adolescent males. We
did not find evidence for adolescent hypersensitivity to reward
value when the reward value was intact in either test set, as
adolescents did not respond significantly more than adults during
the valued condition. Furthermore, all rats reduced their lever
pressing behavior when the outcome was devalued compared with
valued, suggesting that adolescents and adults are equally sensitive
to a change in relative reward value. As reward value sensitivity is
only one aspect that may contribute to reward sensitivity, future
research is needed to investigate age and sex differences in other
aspects of reward sensitivity, such as Pavlovian-to-instrumental
transfer and progressive ratio responding to measure motivation.
In risk-based decision making, adolescents chose the large option
less than adults when the probability of large reinforcement
was 100%. However, the choices of adolescents and females
during the free choice trials led to higher reinforcement rate per
body weight in these groups compared with their adult and male
counterparts. Despite this initial difference, adolescents and adults
showed similar declining percent choices of the large, risky option
during subsequent sessions with decreasing probabilities of large
reinforcer delivery, suggesting that rats of both ages and sexes
were sensitive to the change in reinforcement probability. However,
all rats, regardless of age or sex, continued to choose the
large, risky option more than 50% of the time even at the lowest
reinforcement probability tested (16.7%), despite the fact that the
more advantageous choice was the small, certain option with this
low probability of large reinforcement. We did not find evidence
for adolescent rats making more risky choices compared to adults,
which conflicts with a previous study showing adolescent males
exhibit increased risky choices (Zoratto et al., 2013).
As suggested above, procedural variables may explain these
discrepant findings. Zoratto, Laviola, and Adriani (2013) investigated
risky choice using a home cage probabilistic delivery paradigm.
Rats in their study were pair-housed in a cage partitioned
with a sliding door that was only closed during testing sessions.
Because this task occurred in the home cage where rats were
socially housed, there may have been a social influence on decision
making (Zoratto et al., 2013). Notably, a study in humans
showed that the presence of peers increased risk taking in adolescents
but not adults (Gardner & Steinberg, 2005). In addition, the
food restriction method employed by Zoratto et al. (2013) may
have affected risky choice in a different manner than the minimal
food restriction used in the present study. Previous studies in adult
males showed that the probability discounting curve is less steep in
free-fed rats compared with those adults restricted to 85% of
their free-feeding weight (St. Onge & Floresco, 2009). Given these
points, future studies of risk taking behavior in rodent models of
adolescence should examine these methodological factors as potential
determinants of age differences in risky behavior.
Influential theories of adolescent vulnerability (Casey et al.,
2011; Shulman et al., 2015; Steinberg, 2010) propose that some of
the behavioral traits adolescents often express, including poor
cognitive flexibility, hypersensitivity to reward, and the tendency
to behave in a risky manner, contribute to the adolescent’s heightened
susceptibility to psychopathologies. The present findings in a
rodent model of adolescence are inconsistent with a straightforward
interpretation of these theories and instead highlight the
importance of factors such as pubertal status and motivational
state. The onset of puberty is a notable adolescent event that occurs
earlier in females compared with males. The experimental design
of the present study did not allow for investigation of males and
females at comparable pubertal stages, which may have impacted
our ability to detect sex differences in the adolescent group. Future
investigations are needed, as only recently have rodent studies of
adolescent behavior begun to accommodate for the differential
pubertal timing between sexes (Willing et al., 2016). In addition,
there is a growing interest in human studies to examine pubertal
development as a factor in relation to adolescent neurobehavioral
development, especially in domains of reward processing and
cognitive control.
A recent longitudinal study (Braams, van Duijvenvoorde, Peper,
& Crone, 2015) found an adolescent peak in nucleus accumbens
activity in response to reward similar to that reported previously
(Galvan et al., 2006). Braams et al. (2015) extended previous
findings by showing that pubertal development was linearly associated
with the nucleus accumbens activity, such that the greatest
nucleus accumbens response occurred during midpuberty. Testosterone
levels were also positively correlated with nucleus accumbens
activity during adolescence, suggesting that pubertal hormones
may play a role in the heightened reward response (Braams
et al., 2015). In addition, sensation seeking peaks during adolescence
in males but not females (Steinberg et al., 2008) and is
positively correlated with pubertal development even when controlling
for age (Martin et al., 2002; Steinberg et al., 2008),
suggesting that pubertal status and age can have dissociable contributions
to adolescent reward behavior.
In contrast to heightened reward system activation, the extent of
orbital frontal cortex activation in adolescents was also more
comparable with children than adults (Galvan et al., 2006). These
results are consistent with a relatively immature cognitive control
system. An overactive reward system and reduced cognitive control
system is proposed to result in increased adolescent risk
taking, which seems to be dependent on sex. Consistent with the
peak in sensation seeking in adolescent males (Steinberg et al.,
2008), self-report of risk taking behavior was associated with
males, and positively correlated with pubertal development
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
AGE AND SEX DIFFERENCES IN INSTRUMENTAL BEHAVIOR 85
(Collado-Rodriguez, MacPherson, Kurdziel, Rosenberg, & Lejuez,
2014). Other studies of cognitive control have shown that cognitive
flexibility increases across adolescence in males with females
displaying more flexible behavior than males earlier in adolescence
(Kalkut et al., 2009). Overall, the human literature suggests
that adolescent hypersensitivity to reward may be tied to pubertal
development, with important sex differences in the development of
reward processing and cognitive control behaviors.
In conclusion, the present study found that adolescent rats earn
higher reinforcement rates than adults under minimal food restriction.
Although this result is consistent with our a priori predictions
that adolescents would be hypersensitive to reward, our remaining
findings were contrary to expectations. Namely, adolescents exhibited
greater behavioral flexibility and had similar risk preference
and sensitivity to reward value compared to adults. Thus,
most of our findings were not in line with theories of adolescent
vulnerability that have been largely supported by studies in humans.
Importantly, the present study suggests that adolescent
behaviors that may confer vulnerability to psychopathology are not
evident under all circumstances in rodent models. Rather our
findings and that of the broader rodent literature in adolescent
behavioral development speak to the influence of factors, such as
pubertal status and motivational state in these discrepancies. Future
work is needed to investigate the methodological, and ultimately
neural, determinants of adolescent-typical behaviors in
rodent models.
References
Andrzejewski, M. E., Schochet, T. L., Feit, E. C., Harris, R., McKee, B. L.,
& Kelley, A. E. (2011). A comparison of adult and adolescent rat
behavior in operant learning, extinction, and behavioral inhibition paradigms.
Behavioral Neuroscience, 125, 93–105. http://dx.doi.org/10
.1037/a0022038
Anker, J. J., Baron, T. R., Zlebnik, N. E., & Carroll, M. E. (2012).
Escalation of methamphetamine self-administration in adolescent and
adult rats. Drug and Alcohol Dependence, 124, 149–153. http://dx.doi
.org/10.1016/j.drugalcdep.2012.01.004
Anker, J. J., & Carroll, M. E. (2010). Reinstatement of cocaine seeking
induced by drugs, cues, and stress in adolescent and adult rats. Psychopharmacology,
208, 211–222. http://dx.doi.org/10.1007/s00213-009-
1721-2
Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery
rate: A Practical and powerful approach to multiple testing. Journal of
the Royal Statistical Society Series B: Methodological, 57, 289–300.
Braams, B. R., van Duijvenvoorde, A. C. K., Peper, J. S., & Crone, E. A.
(2015). Longitudinal changes in adolescent risk-taking: A comprehensive
study of neural responses to rewards, pubertal development, and
risk-taking behavior. The Journal of Neuroscience, 35, 7226–7238.
http://dx.doi.org/10.1523/JNEUROSCI.4764-14.2015
Brown, V. J., & Tait, D. S. (2014). Behavioral flexibility: Attentional
shifting, rule switching, and response reversal. In I. Stolerman & L. Price
(Eds.), Encyclopedia of Psychopharmacology, 209–213.
Casey, B., Jones, R. M., & Somerville, L. H. (2011). Braking and accelerating
of the adolescent brain. Journal of Research on Adolescence, 21,
21–33. http://dx.doi.org/10.1111/j.1532-7795.2010.00712.x
Castellano, J. M., Bentsen, A. H., Sánchez-Garrido, M. A., Ruiz-Pino, F.,
Romero, M., Garcia-Galiano, D., . . . Tena-Sempere, M. (2011). Early
metabolic programming of puberty onset: Impact of changes in postnatal
feeding and rearing conditions on the timing of puberty and development
of the hypothalamic kisspeptin system. Endocrinology, 152, 3396–3408.
http://dx.doi.org/10.1210/en.2010-1415
Cauffman, E., Shulman, E. P., Steinberg, L., Claus, E., Banich, M. T.,
Graham, S., & Woolard, J. (2010). Age differences in affective decision
making as indexed by performance on the Iowa Gambling Task. Developmental
Psychology, 46, 193–207. http://dx.doi.org/10.1037/a0016128
Chen, C. Y., Storr, C. L., & Anthony, J. C. (2009). Early-onset drug use
and risk for drug dependence problems. Addictive Behaviors, 34, 319–
322. http://dx.doi.org/10.1016/j.addbeh.2008.10.021
Collado-Rodriguez, A., MacPherson, L., Kurdziel, G., Rosenberg, L. A., &
Lejuez, C. W. (2014). The relationship between puberty and risk taking
in the real world and in the laboratory. Personality and Individual
Differences, 68, 143–148. http://dx.doi.org/10.1016/j.paid.2014.04.019
Cox, B. M., Cope, Z. A., Parsegian, A., Floresco, S. B., Aston-Jones, G.,
& See, R. E. (2016). Chronic methamphetamine self-administration
alters cognitive flexibility in male rats. Psychopharmacology, 233,
2319–2327. http://dx.doi.org/10.1007/s00213-016-4283-0
Dalton, G. L., Wang, N. Y., Phillips, A. G., & Floresco, S. B. (2016).
Multifaceted contributions by different regions of the orbitofrontal and
medial prefrontal cortex to probabilistic reversal learning. The Journal of
Neuroscience, 36, 1996–2006. http://dx.doi.org/10.1523/JNEUROSCI
.3366-15.2016
Drzewiecki, C. M., Willing, J., & Juraska, J. M. (2016). Synaptic number
changes in the medial prefrontal cortex across adolescence in male and
female rats: A role for pubertal onset. Synapse, 70, 361–368. http://dx
.doi.org/10.1002/syn.21909
Eshel, N., Nelson, E. E., Blair, R. J., Pine, D. S., & Ernst, M. (2007).
Neural substrates of choice selection in adults and adolescents: Development
of the ventrolateral prefrontal and anterior cingulate cortices.
Neuropsychologia, 45, 1270 –1279. http://dx.doi.org/10.1016/j
.neuropsychologia.2006.10.004
Floresco, S. B., Block, A. E., & Tse, M. T. L. (2008). Inactivation of the
medial prefrontal cortex of the rat impairs strategy set-shifting, but not
reversal learning, using a novel, automated procedure. Behavioural
Brain Research, 190, 85–96. http://dx.doi.org/10.1016/j.bbr.2008.02
.008
Friemel, C. M., Spanagel, R., & Schneider, M. (2010). Reward sensitivity
for a palatable food reward peaks during pubertal developmental in rats.
Frontiers in Behavioral Neuroscience, 4, 1–10.
Galvan, A., Hare, T. A., Parra, C. E., Penn, J., Voss, H., Glover, G., &
Casey, B. J. (2006). Earlier development of the accumbens relative to
orbitofrontal cortex might underlie risk-taking behavior in adolescents.
The Journal of Neuroscience, 26, 6885–6892. http://dx.doi.org/10.1523/
JNEUROSCI.1062-06.2006
Gardner, M., & Steinberg, L. (2005). Peer influence on risk taking, risk
preference, and risky decision making in adolescence and adulthood: An
experimental study. Developmental Psychology, 41, 625–635. http://dx
.doi.org/10.1037/0012-1649.41.4.625
Ghods-Sharifi, S., Haluk, D. M., & Floresco, S. B. (2008). Differential
effects of inactivation of the orbitofrontal cortex on strategy set-shifting
and reversal learning. Neurobiology of Learning and Memory, 89, 567–
573. http://dx.doi.org/10.1016/j.nlm.2007.10.007
Green, J. L., Dykstra, L. A., & Carelli, R. M. (2015). Examination of
cocaine dose in a preclinical model of natural reward devaluation by
cocaine. Behavioural Pharmacology, 26, 398–402. http://dx.doi.org/10
.1097/FBP.0000000000000131
Hammerslag, L. R., & Gulley, J. M. (2014). Age and sex differences in
reward behavior in adolescent and adult rats. Developmental Psychobiology,
56, 611–621. http://dx.doi.org/10.1002/dev.21127
Hankosky, E. R., Westbrook, S. R., Haake, R. M., Marinelli, M., & Gulley,
J. M. (2017). Reduced sensitivity to reinforcement in adolescent compared
to adult Sprague-Dawley rats of both sexes. Psychopharmacology.
Advance online publication. http://dx.doi.org/10.1007/s00213-017-
4804-5
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
86 WESTBROOK, HANKOSKY, DWYER, AND GULLEY
Jupp, B., & Dalley, J. W. (2014). Behavioral endophenotypes of drug
addiction: Etiological insights from neuroimaging studies. Neuropharmacology,
76(PART B), 487–97.
Kalkut, E. L., Han, S. D., Lansing, A. E., Holdnack, J. A., & Delis, D. C.
(2009). Development of set-shifting ability from late childhood through
early adulthood. Archives of Clinical Neuropsychology, 24, 565–574.
http://dx.doi.org/10.1093/arclin/acp048
Kantak, K. M., Goodrich, C. M., & Uribe, V. (2007). Influence of sex,
estrous cycle, and drug-onset age on cocaine self-administration in rats
(Rattus norvegicus). Experimental and Clinical Psychopharmacology,
15, 37–47. http://dx.doi.org/10.1037/1064-1297.15.1.37
Kim, Y., Simon, N. W., Wood, J., & Moghaddam, B. (2016). Reward
anticipation is encoded differently by adolescent ventral tegmental area
neurons. Biological Psychiatry, 79, 878–886. http://dx.doi.org/10.1016/
j.biopsych.2015.04.026
Korenbrot, C. C., Huhtaniemi, I. T., & Weiner, R. I. (1977). Preputial
separation as an external sign of pubertal development in the male rat.
Biology of Reproduction, 17, 298 –303. http://dx.doi.org/10.1095/
biolreprod17.2.298
Kumar, G., Talpos, J., & Steckler, T. (2015). Strain-dependent effects on
acquisition and reversal of visual and spatial tasks in a rat touchscreen
battery of cognition. Physiology & Behavior, 144, 26–36. http://dx.doi
.org/10.1016/j.physbeh.2015.03.001
Markham, J. A., Morris, J. R., & Juraska, J. M. (2007). Neuron number
decreases in the rat ventral, but not dorsal, medial prefrontal cortex
between adolescence and adulthood. Neuroscience, 144, 961–968. http://
dx.doi.org/10.1016/j.neuroscience.2006.10.015
Marshall, A. T., Liu, A. T., Murphy, N. P., Maidment, N. T., & Ostlund,
S. B. (2017). Sex-specific enhancement of palatability-driven feeding in
adolescent rats. PLoS ONE, 12(7), e0180907. http://dx.doi.org/10.1371/
journal.pone.0180907
Martin, C. A., Kelly, T. H., Rayens, M. K., Brogli, B. R., Brenzel, A.,
Smith, W. J., & Omar, H. A. (2002). Sensation seeking, puberty, and
nicotine, alcohol, and marijuana use in adolescence. Journal of the
American Academy of Child & Adolescent Psychiatry, 41, 1495–1502.
http://dx.doi.org/10.1097/00004583-200212000-00022
Mitchell, M. R., Weiss, V. G., Beas, B. S., Morgan, D., Bizon, J. L., &
Setlow, B. (2014). Adolescent risk taking, cocaine self-administration,
and striatal dopamine signaling. Neuropsychopharmacology, 39, 955–
962. http://dx.doi.org/10.1038/npp.2013.295
Naneix, F., Marchand, A. R., Di Scala, G., Pape, J.-R., & Coutureau, E.
(2012). Parallel maturation of goal-directed behavior and dopaminergic
systems during adolescence. The Journal of Neuroscience, 32, 16223–
16232. http://dx.doi.org/10.1523/JNEUROSCI.3080-12.2012
National Research Council. (2011). Guide for the care and use of laboratory
animals. Washington, DC: National Academies Press.
Newman, L. A., & McGaughy, J. (2011). Adolescent rats show cognitive
rigidity in a test of attentional set shifting. Developmental Psychobiology,
53, 391–401. http://dx.doi.org/10.1002/dev.20537
Orsini, C. A., & Setlow, B. (2017). Sex differences in animal models of
decision making. Journal of Neuroscience Research, 95, 260–269.
http://dx.doi.org/10.1002/jnr.23810
Quinn, J. J., Hitchcott, P. K., Umeda, E. A., Arnold, A. P., & Taylor, J. R.
(2007). Sex chromosome complement regulates habit formation. Nature
Neuroscience, 10, 1398–1400. http://dx.doi.org/10.1038/nn1994
Shahbazi, M., Moffett, A. M., Williams, B. F., & Frantz, K. J. (2008). Ageand
sex-dependent amphetamine self-administration in rats. Psychopharmacology,
196, 71–81. http://dx.doi.org/10.1007/s00213-007-0933-6
Shram, M. J., Funk, D., Li, Z., & Lê, A. D. (2008). Nicotine selfadministration,
extinction responding and reinstatement in adolescent
and adult male rats: Evidence against a biological vulnerability to
nicotine addiction during adolescence. Neuropsychopharmacology, 33,
739–748. http://dx.doi.org/10.1038/sj.npp.1301454
Shulman, E. P., Harden, K. P., Chein, J. M., & Steinberg, L. (2015). Sex
differences in the developmental trajectories of impulse control and
sensation-seeking from early adolescence to early adulthood. Journal of
Youth and Adolescence, 44, 1–17. http://dx.doi.org/10.1007/s10964-
014-0116-9
Simon, N. W., Gregory, T. A., Wood, J., & Moghaddam, B. (2013).
Differences in response initiation and behavioral flexibility between
adolescent and adult rats. Behavioral Neuroscience, 127, 23–32. http://
dx.doi.org/10.1037/a0031328
Steinberg, L. (2008). A social neuroscience perspective on adolescent
risk-taking. Developmental Review, 28, 78–106. http://dx.doi.org/10
.1016/j.dr.2007.08.002
Steinberg, L. (2010). A behavioral scientist looks at the science of adolescent
brain development. Brain and Cognition, 72, 160–164. http://dx
.doi.org/10.1016/j.bandc.2009.11.003
Steinberg, L., Albert, D., Cauffman, E., Banich, M., Graham, S., &
Woolard, J. (2008). Age differences in sensation seeking and impulsivity
as indexed by behavior and self-report: Evidence for a dual systems
model. Developmental Psychology, 44, 1764–1778. http://dx.doi.org/10
.1037/a0012955
St. Onge, J. R., & Floresco, S. B. (2009). Dopaminergic modulation of
risk-based decision making. Neuropsychopharmacology, 34, 681–697.
http://dx.doi.org/10.1038/npp.2008.121
Sturman, D. A., Mandell, D. R., & Moghaddam, B. (2010). Adolescents
exhibit behavioral differences from adults during instrumental learning
and extinction. Behavioral Neuroscience, 124, 16–25. http://dx.doi.org/
10.1037/a0018463
Sturman, D. A., & Moghaddam, B. (2011). Reduced neuronal inhibition
and coordination of adolescent prefrontal cortex during motivated behavior.
The Journal of Neuroscience, 31, 1471–1478. http://dx.doi.org/
10.1523/JNEUROSCI.4210-10.2011
Van Leijenhorst, L., Westenberg, P. M., & Crone, E. A. (2008). A developmental
study of risky decisions on the cake gambling task: Age and
gender analyses of probability estimation and reward evaluation.
Developmental Neuropsychology, 33, 179 –196. http://dx.doi.org/10
.1080/87565640701884287
Willing, J., Drzewiecki, C. M., Cuenod, B. A., Cortes, L. R., & Juraska,
J. M. (2016). A role for puberty in water maze performance in male and
female rats. Behavioral Neuroscience, 130, 422–427. http://dx.doi.org/
10.1037/bne0000145
Willing, J., & Juraska, J. M. (2015). The timing of neuronal loss across
adolescence in the medial prefrontal cortex of male and female rats.
Neuroscience, 301, 268–275. http://dx.doi.org/10.1016/j.neuroscience
.2015.05.073
Wong, W. C., Ford, K. A., Pagels, N. E., McCutcheon, J. E., & Marinelli,
M. (2013). Adolescents are more vulnerable to cocaine addiction: Behavioral
and electrophysiological evidence. The Journal of Neuroscience,
33, 4913–4922. http://dx.doi.org/10.1523/JNEUROSCI.1371-12
.2013
Zoratto, F., Laviola, G., & Adriani, W. (2013). Gambling proneness in rats
during the transition from adolescence to young adulthood: A homecage
method. Neuropharmacology, 67, 444–454. http://dx.doi.org/10
.1016/j.neuropharm.2012.11.024
Received September 9, 2017
Revision received December 20, 2017
Accepted January 11, 2018
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
AGE AND SEX DIFFERENCES IN INSTRUMENTAL BEHAVIOR 87

Put Your Full Paper Title Here in Title Case

Your Name

Department, School

Course number and name

Dr. Sabine Huemer

Date

Put Your Full Paper Title Here in Title Case

What were the variables examined? Identify them as IV and DV (note that there can be more than one IV or DV). If you get confused, remember to formulate the sentence: The effects of (IV) on (DV); often you find the main variables in the title of the article, in the Abstract, and at the end of the Introduction section.

What was the hypothesis (there can be more than one)? You will find this at the end of the Introduction (NOT the Abstract) just before the Methods section.

Materials and Methods

Describe the participants in the study. If there was more than one study, write them up for each study: How many participants were there in total, age range, gender breakdown, any other essential information on participant demographics? Describe the sampling method. This information will be found in the Methods section.

Describe the MATERIALS used to test the hypothesis. Note that there is a lot of testing in experiments, some of the testing may not be related to hypothesis testing rather to trying to avoid confounds. Reread the hypothesis and zoom in on how they tested them: name the tests they used and tell us how they manipulated the IV (what groups/treatment conditions did they create and how were they different from each other). This is part of the Methods section.

Please describe what study design was used: Tell us if it was an experiment or quasi-experiment, then add whether it was a between-subject, within-subject, matched-subject, or, factorial design. The latter may not be explicitly stated but I want you to figure it out based on the information given.

Note that in a quasi-experiment, the group assignment is determined by a personal characteristic and it is not the choice of the researcher, i.e.: if it is a study on ADHD with a control group, all people with ADHD will be in one group and the control group will be people without ADHD so there cannot be random assignment. In a factorial design, there is more than one independent variable.

Describe the procedure of the experiment which is the chronological sequence of events. Only put in what is essential to understand how the experiment was conducted. This is part of the Methods section.

Results

Report the results of the hypothesis testing. Do not get sidetracked by results from testing that did not relate to the hypothesis. Note that the detailed results can be found in the Results section. This section often becomes overwhelming if you are not used to reading them. In that case, go to the Abstract and also the first portion of the Discussion section to find a simple explanation of what was found. It will be sufficient to summarize the results in that form. If you feel more confident, add some numbers from the results section. Look for statistical significance.

Discussion

Critically evaluate the experiment: What are the strength and the weaknesses? The study will discuss this in the Discussion section but you can add your own critical thoughts here. What were the limitations of the study? Why is this study important and who would benefit from the results and how? How can future studies build on the results from this study?

References

Lastname, I. (2010). Title of article with only first word capped or any proper noun. Title of Publication, 22(3), 325-340. doi or URL.