Cognitive Training Can Reduce Civilian Casualties in a Simulated Shooting Environment

Abstract

Shooting a firearm involves a complex series of cognitive abilities. For example, locating an item or a person of interest requires visual search, and firing the weapon (or withholding a trigger squeeze) involves response execution (or inhibition). The present study used a simulated shooting environment to establish a relationship between a particular cognitive ability and a critical shooting error-response inhibition and firing on civilians, respectively. Individual-difference measures demonstrated, perhaps counterintuitively, that simulated civilian casualties were not related to motor impulsivity (i.e., an itchy trigger finger) but rather to an individual's cognitive ability to withhold an already initiated response (i.e., an itchy brain). Furthermore, active-response-inhibition training reduced simulated civilian casualties, which revealed a causal relationship. This study therefore illustrates the potential of using cognitive training to possibly improve shooting performance, which might ultimately provide insight for military and law-enforcement personnel.
© The Author(s) 2015.

Full text

Psychological Science
2015, Vol. 26(8) 1164 –1176
© The Author(s) 2015
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0956797615579274
pss.sagepub.com
Research Article
Shooting a firearm involves a complex cascade of actions,
and each action can be linked to a specific cognitive abil-
ity. For example, finding appropriate targets involves
visual search, determining whether someone is a friend
or foe involves decision-making processes, taking aim
involves perceptual estimations of distance and motion,
and squeezing the trigger (or withholding the shot)
involves response execution (or inhibition). Given this
hypothesized series of cognitive functions, shooting
behaviors provide an excellent opportunity to examine
the links between action and cognition. However, there
is little existing evidence linking shooting performance
and cognition, which is why, in the current study, we
attempted to support this larger concept by providing
initial evidence to tie a particular shooting error to a par-
ticular cognitive ability.
The present investigation focused on one potential
relationship between shooting and cognition—civilian
casualties and response inhibition. Civilian casualties
occur when shooters hit noncombatants with weapons
fire (Kahl, 2007; Wright, 2003; cf. friendly-fire incidents,
in which an ally is hit with weapons fire; Webb & Hewett,
2010), and this critical shooting error can have dramatic
psychological, ethical, economic, and practical implica-
tions. Thus, every effort should be made to minimize
these occurrences. In the current project, we looked to
understand and provide potential insight into reducing
civilian casualties by comparing performance in a simu-
lated shooting environment with response inhibition—
the ability to stop performing an already initiated behavior
(Eagle et al., 2008; Logan, 1994; Logan, Schachar, &
Tannock, 1997; Menon, Adleman, White, Glover, & Reiss,
2001; Verbruggen & Logan, 2008). For example, when
579274PSSXXX10.1177/0956797615579274Biggs et al.Reducing Civilian Casualties
research-article2015
Corresponding Author:
Adam T. Biggs, Duke University, Center for Cognitive Neuroscience,
B203 Levine Science Research Center, Box 90999, Durham, NC 27708
E-mails: adam.t.biggs@gmail.com, adam.biggs@duke.edu
Cognitive Training Can Reduce Civilian
Casualties in a Simulated Shooting
Environment
Adam T. Biggs1, Matthew S. Cain2, and Stephen R. Mitroff1
1Center for Cognitive Neuroscience, Duke University, and 2U.S. Army Natick Soldier Research,
Development, & Engineering Center, Natick, Massachusetts
Abstract
Shooting a firearm involves a complex series of cognitive abilities. For example, locating an item or a person of interest
requires visual search, and firing the weapon (or withholding a trigger squeeze) involves response execution (or
inhibition). The present study used a simulated shooting environment to establish a relationship between a particular
cognitive ability and a critical shooting error—response inhibition and firing on civilians, respectively. Individual-
difference measures demonstrated, perhaps counterintuitively, that simulated civilian casualties were not related to
motor impulsivity (i.e., an itchy trigger finger) but rather to an individual’s cognitive ability to withhold an already
initiated response (i.e., an itchy brain). Furthermore, active-response-inhibition training reduced simulated civilian
casualties, which revealed a causal relationship. This study therefore illustrates the potential of using cognitive training
to possibly improve shooting performance, which might ultimately provide insight for military and law-enforcement
personnel.
Keywords
shooting cognition, guns, attention, response inhibition, cognitive training, civilian casualties
Received 10/12/14; Revision accepted 3/5/15
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

Reducing Civilian Casualties 1165
you start to hit the “send” button on an e-mail, you might
suddenly realize that it is addressed to the wrong person.
Sometimes you abort the button-press behavior (i.e., suc-
cessful response inhibition), but sometimes you do not
(i.e., failed response inhibition). Initiated responses can
be successfully withheld despite very brief time windows
between the decision to respond and the response-
inhibition signal (Bissett & Logan, 2011; Chikazoe etal.,
2009; Leotti & Wager, 2010), although this process can be
cognitively challenging. When applying the logic of
response inhibition to shooting a firearm, consider the
situation in which shooters initiate the process to fire the
weapon—but then realize that their target is a civilian. In
this case, the shooters must rely on response-inhibition
skills to successfully inhibit a trigger squeeze.
Exploring this particular relationship between shoot-
ing and cognition can offer numerous practical benefits.
First, the action-cognition links between shooting and
cognitive abilities could help identify individuals well-
suited for performing specific shooting tasks—or con-
versely, individuals who are most likely to make
particular types of errors. For example, if the proposed
relationship between civilian casualties and response
inhibition exists, the individuals most likely to inflict
civilian casualties could be identified before being sent
into combat. Second, if specific cognitive abilities can
predict specific aspects of shooting performance, then
individualized training could be developed to help indi-
viduals avoid the particular errors to which they are pre-
disposed. Ultimately, these combined efforts could
potentially inform military and law-enforcement efforts
in performance, training, and evaluation. Such new
training methods are particularly important given recent
evidence that deliberate practice alone may not be as
influential for performance in professional tasks as once
believed (Macnamara, Hambrick, & Oswald, 2014). That
said, new training methods do not dismiss the impor-
tance of practice in improving performance; rather, they
highlight the potential in improving performance beyond
practice alone.
The present study contained three components to test
this proposed relationship: baseline participants to assess
relationships between attentional abilities and simulated
civilian casualties, a response-inhibition training (RIT)
group to assess the efficacy of training, and an active con-
trol training group (visual search training, or VST). Baseline
measurements consisted of shooting performance, cogni-
tive abilities, and self-report surveys to examine whether
simulated civilian casualties were related to individual dif-
ferences in response inhibition, attentional deficits, or
impulsivity. RIT consisted of three sessions with a 30-min
computer-based stop-signal reaction time (SSRT) task and
a 30-min iPad-based go/no-go task. VST consisted of three
sessions with a 1-hr, computer-based visual search task
designed to enhance search consistency (Biggs, Cain,
Clark, Darling, & Mitroff, 2013; Biggs & Mitroff, 2014).
Both training groups completed a 5-day protocol: a 2-hr
pretraining session on Day 1, a 1-hr cognitive-training
session on Days 2 through 4, and a 2-hr posttraining ses-
sion on Day 5.
The key dependent variable was civilian casualties
within a simulated shooting environment that contained
both intended targets (i.e., hostile individuals) and unin-
tended targets (i.e., civilians; see Fig. 1). This simulated
shooting scenario was designed to model the key compo-
nents of realistic shooting, including squeezing a trigger
to fire and moving the mock firearm in real space to aim.
We hypothesized that response inhibition would be
related to the number of simulated civilian casualties
inflicted, and therefore training response-inhibition abili-
ties would reduce these unintended casualties. Notably,
cognitive-training designs can be subject to various meth-
odological concerns, such as placebo effects (Boot,
Simons, Stothart, & Stutts, 2013; Green, Strobach, &
Schubert, 2014; Stothart, Simons, Boot, & Kramer, 2014),
which is why we used an experimental protocol with two
active training groups. Visual search, while possibly
related to overall shooting performance, is not conceptu-
ally related to this particular shooting error. As such, VST
should not reduce simulated civilian casualties, but this
active training reduces methodological concerns about
placebo effects because both groups are actively training—
albeit through conceptually different procedures.
Method
Participants
All participants (N = 88; mean age = 24.92 years, SD =
7.48; 52 female, 36 male) completed the baseline session.
A subset (n = 57) was randomly assigned to either the
RIT (28 participants) or VST (29 participants) condition.
The number of participants was determined prior to col-
lecting data and was based on a reasonable number for
individual-difference analyses (approximately 30 per
training group and more than 80 participants for baseline
analyses). We stopped collecting data when we reached
participant numbers that satisfied these criteria.
Procedure
All participants completed a 2-hr baseline (pretraining)
session on Day 1, in which they played the shooting
game and completed five surveys and four computer-
based tasks. In addition, the two training groups com-
pleted a 1-hr cognitive-training session on Days 2 through
4 and a 2-hr posttraining session on Day 5, in which they
again played the shooting game.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

1166 Biggs et al.
Fig. 1. Sample screen shots and equipment from the video game Reload: Target Down for the Nintendo
Wii (Mastiff, 2013; reprinted with permission). The top row provides examples of (a) an instructional
screen showing which civilian targets to avoid, (b) a remote used to simulate firing a weapon during the
experiment, and (c) an instructional screen showing which hostile targets to shoot. The two experimental
scenarios—(d) “Embassy Training” and (e) “Apartment Training”—included both unarmed civilians and
hostile individuals pointing weapons.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

Reducing Civilian Casualties 1167
Baseline session
Surveys. Five self-report surveys were administered
to participants. The Jasper-Goldberg Adult ADD Ques-
tionnaire (Jasper & Goldberg, 1993) assesses attention-
deficit/hyperactivity disorder (ADHD) symptoms, with
higher scores indicating more symptoms. The Barratt
Impulsivity Scale (Barratt, 1959; Patton, Stanford, & Bar-
ratt, 1995) uses three subscales to measure various forms
of impulsivity: attentional impulsivity, motor impulsiv-
ity, and nonplanning impulsivity. The Autism-Spectrum
Quotient (Baron-Cohen, Wheelwright, Skinner, Mar-
tin, & Clubley, 2001) measures autism symptoms, with
higher scores indicating more symptoms. The Maximiza-
tion Scale (Schwartz etal., 2002) assesses an individual’s
desire to maximize outcomes. Finally, we administered
a video-game questionnaire constructed in the Duke
Visual Cognition Lab (see Appelbaum, Cain, Darling, &
Mitroff, 2013), which measured self-reported expertise,
preferences, and extent of video-game playing. Note that
survey data were excluded only if participants did not
complete all questions for that particular scale.
Simulated shooting scenarios. All shooting scenarios
were completed on the Nintendo Wii game Reload: Tar-
get Down (Mastiff, 2013). Participants stood 1.75 meters
away from a 28-in. LCD television screen. A black Wii
Motion Plus remote was placed into a black plastic holder
designed to resemble a more realistic weapon (see Fig.
1b). Participants simulated shooting a firearm by squeez-
ing the trigger, which caused the remote control to vibrate
as identification that the shot had been fired. A targeting
reticle appeared on screen to indicate precisely where
the gun was aimed, and a shot would land precisely at
the center of the crosshairs after a trigger squeeze. This
aspect of game play provided the opportunity to explore
shooting cognition while minimizing concerns about
whether the participant could accurately aim.
All participants began with a practice round (on a sim-
ulated shooting range) to become accustomed to game
play. The simulated weapon for the entire round was a
semiautomatic pistol. Different targets appeared (e.g.,
paper bull’s-eyes, silhouettes, bottles), and the game
slowly introduced various elements to participants (e.g.,
the timer indicating how much time remained in the
round). Each shot could earn a maximum of 100 points
depending on how accurately the participant hit each
simulated target. Targets burst apart on being shot, and a
number appeared in white above the target to indicate
how many points had been earned. Some targets in later
practice rounds required a “double tap” to destroy, which
required that participants hit the target twice in rapid
succession—a procedure that helped participants become
accustomed to firing multiple shots in short order. Points
could also be earned by completing a round (i.e.,
destroying all targets) with time remaining; 100 points
were awarded for every second remaining on the timer.
This encouraged participants to shoot all targets as
quickly as possible in addition to as accurately as possi-
ble. Participants were instructed to gain as many points
as they could during the round, although no participant
was allowed to advance beyond the practice round with-
out successfully destroying 100 targets. Five out of 88
participants failed to meet this 100-target minimum, but
reached at least the minimum after two practice rounds.
The rounds used for experimental data were the sce-
narios entitled “Embassy Training” and “Apartment
Training.” Each scenario presented the same mix of
intended targets (i.e., “bad guys”) and unintended targets
(i.e., “civilians”). Participants were instructed to clear
individual rooms by shooting hostile targets without hit-
ting civilians. Each scenario continued until the partici-
pant had gone through all rooms or failed the scenario.
A participant failed by shooting five civilians, whereupon
the game exited the scenario to a screen that said “You
failed! You have shot too many civilians. You must be
more careful next time.”
Both scenarios presented the same set of four possible
hostile targets and four possible civilians. Individual
rooms in each scenario presented a randomized set of
hostile targets and civilians in a predetermined arrange-
ment of possible positions. All rooms contained at least
one civilian and at least three hostile targets, with a maxi-
mum of five civilians and nine hostile targets. The key
difference between scenarios was the rate of fire for the
weapon (Embassy Training was completed with a semi-
automatic handgun and Apartment Training with a fully
automatic M16 rifle).
All targets were simulated cardboard cutouts that burst
apart when shot. Hostile targets were armed and pointing
weapons at the participant, whereas civilian targets were
unarmed. Points varied based on where the hostile target
had been shot: in the head (100), torso (~80), or legs
(~50). A number in white appeared above the target to
indicate how many points had been earned. Participants
lost 1,000 points for shooting a civilian target, and “−1000”
appeared over the target in red along with an audio mes-
sage (e.g., “Stop shooting hostages!” or “Watch out for
civilians!”).
Participants completed six scenarios counterbalanced
by type to limit any memorization of target positions
(e.g., if the participant went through Embassy Training
first, then the order would be Embassy, Apartment,
Embassy, Apartment, Embassy, Apartment). The first pass
through each scenario was treated as practice as it was
the first experience with the fully automatic rate of fire,
the first experience with hostile and civilian targets, or
potentially both. The first pass-through thus familiarized
participants with an automatic weapon and helped the
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

1168 Biggs et al.
participant dissociate between hostile targets and
civilians— the game heavily reinforced that civilians
should not be shot (the penalty was 10 times larger than
the maximum points earned for shooting a bad guy,
shooters failed the round if they hit five civilians, and the
audio message exhorted shooters not to hit civilians).
Additionally, the first pass yielded the score to beat for
each individual. This gamelike feature provided an incen-
tive and an attainable goal for each subsequent round to
ensure effortful performance (cf. Miranda & Palmer,
2014). The experimental data were drawn from the final
four rounds: Our civilian-casualties measure used
throughout this study was the sum of civilian targets hit
during these four rounds.
Computer-based tasks. Participants completed four
computer-based tasks at baseline: a go/no-go task, an
SSRT task, a Stroop interference task, and a visual search
task (see Fig. 2). Participants completed the computer-
based tasks on Dell Vostro 260 computers with 23.6-in.
widescreen LCD monitors. Stimuli were presented and
responses collected with MATLAB software (The Math-
Works, Natick, MA) and the Psychophysics Toolbox
(Version 3.0.8; Brainard, 1997; Kleiner, Brainard, & Pelli,
2007; Pelli, 1997). Participants were seated approximately
57 cm from the screen without head restraint.
In the go/no-go task, blue and orange squares (1.32°×
1.32°) appeared one at a time against a gray background.
Participants pressed the space bar if one color appeared
(the go signal) and withheld a response if the other color
appeared (the no-go signal). The go-signal color was
counterbalanced across participants. The colors of the go
and no-go signals were reversed at posttraining for par-
ticipants in the two training conditions (e.g., someone
with a blue go signal at pretraining had an orange go
signal at posttraining). Each trial began with a fixation
circle that appeared for a randomly determined time
between 0.5 s and 1.5 s before the stimulus appeared.
The colored square then appeared for up to 2 s at fixa-
tion before the computer proceeded to the next trial. The
200 experimental trials were preceded by 20 practice tri-
als with 80% go signals and 20% no-go signals. No-go
Go/No-Go Task
Stroop Task Visual Search Task
Stop-Signal Reaction Time Task
Fixation Display
(0.5–1.5 s)
Go Signal or No-Go Signal
(Up to 2 s)
Fixation Display
(1.25–1.625 s)
Go Signal Until Response but
Could Change Into Stop Signal
on 1/3 of Trials
Letter String Presented in Green, Red, or Blue
Fixation Display
(0.5 s)
Fig. 2. The four tasks completed in the baseline session. In the go/no-go task, participants pressed the space bar when one of two colors
appeared and withheld response when the other color appeared. The stop-signal reaction time task was similar, except that the trial always
began with a go signal that would sometimes change into a no-go signal. On the Stroop interference task, participants had to identify the font
color of letter strings that contradicted, did not contradict, or had no relation to that color. On the visual search task, participants identified
whether a target was present or absent among distractors in an array.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

Reducing Civilian Casualties 1169
signal accuracy for experimental trials was measured as
the percentage of trials in which the participant correctly
withheld a response.
The SSRT task was similar to the go/no-go task in
structure, but it assessed a different aspect of response-
inhibition abilities: the ability to withhold an already initi-
ated response. Participants responded as quickly as
possible when they saw a go signal, yet some trials
included a change—the go signal turned into a no-go
signal. The task required participants to withhold a
response when they saw the signal change.
Green and purple squares (3.18° × 3.18°) appeared on
screen one at a time against a black background.
Participants were instructed to press the space bar if one
color appeared (the go signal) but withhold a response if
the other color appeared (the stop signal). The go-signal
color was counterbalanced across participants. Each trial
began with a fixation circle that appeared for a randomly
determined time between 1.25 s and 1.625 s. On go trials,
a go signal appeared and remained on screen until
response. On stop trials, a go signal appeared initially,
but the color switched to the stop signal during the trial
and remained on screen for 1 s afterward. Two-thirds of
trials were go trials, and one third were stop trials.
Participants completed a first block of 90 trials as practice
(discarded from analyses) and then experimental trials
until reaching 312 trials or a maximum of 16 min spent
completing the task. Only 1 participant out of 88 reached
the time limit. In the experimental blocks, participants
received a warning beep from the computer if their go-
signal response time (RT) exceeded their mean go-signal
RT from the practice block by more than 2 standard
deviations.
Performance was primarily determined by the stop-
signal delay (SSD)—the time difference between the
appearance of the go signal and its change into a no-go
signal. The SSD was altered on the basis of an individu-
al’s performance by utilizing a one-up/one-down stair-
case procedure. If participants correctly withheld the
response when a stop signal appeared, the SSD increased
by 33 ms to make stopping more difficult. If participants
did not withhold the response when a stop signal
appeared, the SSD was reduced by 33 ms to make stop-
ping easier. The time needed to respond to a go signal
was calculated as the median RT for correct responses on
go-signal trials (no response or hitting a key other than
the space bar would have been an incorrect response).
However, stopping time could not be directly measured,
because no overt response was to be made.
The staircase procedure was designed to produce
50% accuracy on stop-signal trials. Final SSRT for each
participant was calculated using the integration method
(Verbruggen, Aron, Stevens, & Chambers, 2010; Ver-
bruggen, Chambers, & Logan, 2013). This approach uses
the go-signal RT on the nth trial minus the average SSD
to calculate the SSRT, for which the nth trial is calculated
by rank-ordering the correct go-signal RTs and using the
trial that corresponded to the probability of responding
on a stop-signal trial (i.e., failing to withhold a response).
For example, a participant might see 150 go-signal trials
during the experiment and respond accurately on 56%
of the stop-signal trials (i.e., successfully withhold a
response), which would indicate that the nth trial con-
tained the 66th fastest correct go-signal response. SSRT
was calculated for all trials across the experiment.
The experiment-wide SSRT integration-method calcu-
lation could be errant if there is gradual slowing across
the experiment (Verbruggen etal., 2013). However, the
correct go-signal RTs revealed no systematic effect of
slowing (Block 1: M = 473 ms, SD = 161 ms; Block 2: M =
465 ms, SD = 165 ms; Block 3: M = 470 ms, SD = 207 ms).
Note that data from 4 participants were removed: 2 par-
ticipants who had stop-signal accuracy more than 2.5
standard deviations below the group mean and 2 partici-
pants who had average go-signal RTs more than 2.5 stan-
dard deviations above the group mean.
In the Stroop interference task, participants viewed
words printed in red, green, or blue font and were
required to identify the font color of the word. Each letter
was approximately 1.22° × 1.03°. Incompatible trials pre-
sented a printed word that contradicted the font color
(e.g., the word “RED” written in blue font), neutral trials
presented a string of three to five “X”s in any color (i.e.,
red, green, or blue), and compatible trials presented a
word that matched the font (e.g., the word “RED” appear-
ing in red font).
Each trial began with a dot appearing at fixation for
0.5 s. A letter string then appeared, and participants iden-
tified the font color by pressing the left arrow key to
signify red font, the down arrow key to signify green
font, and the right arrow key to signify blue font.
Participants completed 30 practice trials before 180
experimental trials, which were divided between incom-
patible (20%), neutral (20%), and compatible (60%).
Stroop interference effects were measured by taking the
RT for incompatible trials with correct responses and
subtracting the RT for neutral trials with correct responses.
In the visual search task, participants searched among
a display of 32 items for a prespecified target item and
indicated whether the target was present or absent (for
details about the stimuli, see Biggs, Cain, etal., 2013).
Target-absent trials consisted of all “L”-shaped distractors,
whereas target-present trials contained a target “T” among
the distractors. Each trial began with a fixation cross for
250 ms. The cross then disappeared, and the search array
appeared until response. Target-present and target-absent
judgments were made using one of two assigned keys
(“z” and “/”; counterbalanced across participants). Ten
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

1170 Biggs et al.
practice trials preceded 100 experimental trials. Both
practice and experimental trials were divided among
equal numbers of target-present and target-absent trials.
Accuracy feedback was provided for practice trials but
not for experimental trials. Search arrays disappeared
after a response, and the next trial began automatically.
Participants were given the opportunity to rest every 25
experimental trials. Search accuracy was calculated as the
number of trials in which the participant correctly
responded that the target was present (i.e., a hit) or that
the target was absent (i.e., a correct rejection).
Response-inhibition training (RIT). Participants in
the RIT condition completed three 1-hr training sessions.
Each session consisted of two tasks: an SSRT task and an
interactive go/no-go task (see Table 1). Each task took
approximately 30 min, and task difficulty adapted to indi-
vidual performance—that is, the task became more diffi-
cult as the participant performed better.
RIT SSRT task. Previous evidence has demonstrated
that inhibitory control could be trained through stop-
signal tasks (Berkman, Kahn, & Merchant, 2014; Manuel,
Bernasconi, & Spierer, 2013), and so a stop-signal task
was included in RIT. Stimuli were left- or right-pointing
arrows (4.3° × 3.7°) presented at the center of the screen
(see Fig. 3). The same stimulus colors (green and purple)
were used here as in the baseline task, and the go-signal
versus stop-signal color was counterbalanced across par-
ticipants and across training sessions. Many aspects of the
design were similar to the baseline SSRT task, including
the one-up/one-down staircase procedure used to alter
the SSD. However, participants now responded by indi-
cating the direction of the arrow (left or right) as opposed
to simply hitting the space bar for the go signal. A fixa-
tion dot appeared for 1.25 s before being replaced by
the arrow (see Fig. 3). The task ended when participants
completed five blocks of trials with 104 trials per block
or they reached a 30-min time limit (forced ceiling). In
the experimental blocks, participants received a warning
beep from the computer if their go-signal RT exceeded
their mean go-signal RT from the practice block by more
than 2 standard deviations.
On Training Day 1 (Day 2 of the study), participants
responded via pressing either the left or right arrow on
the keyboard with their dominant hand only (i.e., if they
squeezed the trigger with their right hand for the shoot-
ing assessments, then only the right hand could be used
to respond on Training Day 1). On Training Days 2 and
3 (Days 3 and 4 of the study), participants were required
to use both hands to respond (pressing the “z” key for a
left-pointing arrow and the “/” key for a right-pointing
arrow). The one-up/one-down staircase was calculated
separately for each hand as only the left hand was used
to respond to left-pointing arrows, and only the right
hand was used to respond to right-pointing arrows.
Interactive go/no-go task. Participants in the RIT con-
dition also played the iPad game Smack That Gugl! Par-
ticipants smacked puttylike figures (“gugls”) by tapping
the screen. Some gugls required only one tap, some
required two, certain gugls could split from one into two
when tapped (and then participants had to tap the two
new gugls), and some had spikes or red bumps to indi-
cate that they should not be tapped (i.e., they required
response inhibition).
Participants began each level with five lives. A life was
lost if participants failed to tap gugls quickly enough or if
they tapped a gugl with spikes or red bumps. Participants
began each training session at Level 1 and proceeded
until they lost five lives. During the training portion of
each day, participants began at the highest level they had
previously completed and continued until losing five
more lives. Participants advanced to a new level by
smashing 100 gugls (e.g., participants went from Level 2
to Level 3 after smashing 200 gugls and would restart
Table 1. Description of the Two Training Conditions
Characteristic
Response-inhibition training (RIT)
Visual search trainingRIT SSRT task Smack That Gugl! game
Task goal Avoiding civilian casualties Avoiding civilian casualties Detecting targets
Duration 30 min 30 min 1 hr
Task medium Computer Handheld tablet Computer
Stimuli Green and purple arrows Claylike blobs Rotated “C”s
Purpose Learn to withhold an initiated
response
Learn to avoid certain stimuli while
performing quickly
Learn to consistently perform a
visual search
Difficulty Increased with correct responses Increased with accuracy Increased by training day
Task Respond to the go color, withhold
the response for the stop color
“Smash” blobs with finger taps,
avoid smashing some blobs
Make a present or absent judgment
about a reversed “C”
Note: SSRT = stop-signal reaction time.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

Reducing Civilian Casualties 1171
from Level 3 after losing five lives). Each training day
ended with a final round in which participants again
began from Level 1 and proceeded as far into the game
as they could before losing five lives.
Visual search training (VST). Participants in the VST
condition completed three training sessions designed to
enhance search consistency with a particular search strat-
egy. Each session increased the difficulty for consistent
search: Session 1 (on Training Day 2) presented perfectly
aligned grid displays, Session 2 (on Training Day 3) intro-
duced spatial jitter, and Session 3 (on Training Day 4)
introduced spatial jitter and gaps between items (see
Table 1).
Each display item was a circle (0.6° in diameter) with a
portion removed to make it resemble the letter “C” (see
Fig. 3). Targets were perfectly reversed “C”s (i.e., rotated
180°), and distractors were drawn from a pool of “C” stim-
uli rotated in increments of 5° (5°, 10°, 15°, etc., except for
175° and 185°). Each training day included three parts. In
Part 1, full displays were presented, and participants were
required to make one target present/absent response per
display. In Part 2, each display was presented one line at
a time starting from the top, and participants were required
to make a target present/absent decision about each line
of the display before it disappeared and the next line
appeared. Part 3 was identical to Part 1.
On all training days, participants were instructed to
search left to right, starting at the top left as though “read-
ing from a book.” On Training Day 1, search displays
were presented with 56 items perfectly aligned in an 8
(horizontal) × 7 (vertical) grid during Part 1 and with 8
items per line during Part 2. On Training Day 2, search
displays were presented with 56 total display items
aligned in an 8 × 7 grid during Part 1 and with 8 items
per line during Part 2, but randomized spatial jitter was
applied to prevent perfect grid alignment. On Training
Day 3, search displays were presented with the same
Training Day 1 Display Training Day 2 Display Training Day 3 Display
Visual Search Training
Response-Inhibition Training
Fixation Display
(1.25 s)
Go Signal Until Response but Could Change
Into Stop Signal on 1/3 of Trials
a
b
Fig. 3. Sample displays from the (a) stop-signal reaction time task in the response-inhibition training condition
and (b) search task in Part 1 of the visual search training condition. In the stop-signal reaction time task, partici-
pants had to indicate the direction in which the go arrow was pointing and make no response for the no-go arrow.
However, the go arrow turned into the no-go arrow on one third of the trials. Visual search training differed across
the 3 training days: On Day 1, perfectly aligned grid displays were presented. On Day 2, spatial jitter was intro-
duced, and on Day 3, there was spatial jitter and gaps between items. On each day, participants had to identify
whether a target (a “C”) was present or absent.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

1172 Biggs et al.
randomized spatial jitter as on Training Day 2, but with 35
total items (5 per line during Part 2). This design allowed
participants to volitionally use the search strategy (Parts 1
and 3) but also provided a more targeted training oppor-
tunity with the provided visual search strategy (Part 2). We
increased the difficulty of using the left-to-right strategy
each day by making the search grid less symmetrical.
Results
Baseline differences
All baseline measures were compared with the number
of civilian casualties in the Reload: Target Down game.
Table 2 presents the correlation and significance values.
Note that participant counts vary by measure because of
data filtering for behavioral performance (i.e., values
more than 3 standard deviations below the group mean
were removed as outliers) and because not all partici-
pants answered all questions in a particular survey.
Performance on the SSRT task, which measured an indi-
vidual’s ability to withhold an initiated response, was sig-
nificantly related to the number of simulated civilian
casualties at baseline, r(82) = .25, p = .02, with poorer
SSRT performance related to greater simulated civilian
casualties. Two self-report scales were significantly
related to the number of simulated civilian casualties at
baseline: More simulated civilian casualties were related
to higher ADHD scores, r(84) = .21, p < .05, and greater
attentional impulsivity, r(82) = .24, p < .05. Simulated
civilian casualties were not related to self-reported motor
impulsivity, r(83) = .12, p = .27, nor the total number of
shots fired in the simulated shooting task—a behavioral
measure of motor impulsivity, r(86) = .08, p = .44.
Training effects
Simulated civilian casualties were significantly reduced
from pretraining to posttraining for the RIT group but not
for the VST group, F(1, 55) = 4.10, p < .05, ηp2 = .07 (Fig.
4). This reduction could not be explained by a group dif-
ference in simulated civilian casualties at pretraining,
t(55) = 0.69, p = .49, nor by a reduction in the number of
intended targets hit, as RIT participants improved more
than VST participants in number of targets correctly shot,
F(1, 55) = 3.87, p = .05, ηp2 = .07. Finally, more self-
reported ADHD symptoms were associated with a larger
reduction in civilian casualties from pretraining to post-
training for the RIT group but not the VST group (Fig. 5).
General Discussion
The current study revealed several possible links between
civilian casualties in a simulated shooting environment
and the cognitive ability of response inhibition. First,
individuals with lower inhibitory control and higher
attentional impulsivity were more likely to shoot civilians
in the simulated scenarios. Second, significant relation-
ships between simulated civilian casualties and atten-
tional measures, but not between civilian casualties and
motor-impulsivity measures, suggest a cognitive under-
pinning of the relationship—an itchy brain more so than
an itchy trigger finger. Third, response-inhibition training
offers exciting potential to inform future training proce-
dures, which might ultimately reduce unintended casual-
ties. Finally, individuals who self-reported high levels of
ADHD symptoms benefited most from the response-inhi-
bition training—which suggests not only that some peo-
ple benefit more from training than others, but also that
such individuals could be identified prior to training.
These findings provide some enticing preliminary evi-
dence that shooting performance could be linked to cog-
nitive abilities and—potentially—that cognitive training
could enhance shooting performance.
These results add to the mounting evidence that inhib-
itory control can be improved through cognitive training
(e.g., Guerrieri, Nederkoorn, & Jansen, 2012; Thorell,
Lindqvist, Bergman, Bohlin, & Klingberg, 2009; for a
review, see Spierer, Chavan, & Manuel, 2013). Some pre-
vious efforts have demonstrated stimulus-specific
response-inhibition training by reducing alcohol con-
sumption via enhanced response inhibition for alcohol-
related stimuli (Houben, Havermans, Nederkoorn, &
Jansen, 2012; Houben, Nederkoorn, Wiers, & Jansen,
2011), or more generalized response-inhibition training
by demonstrating reduced risk taking in gambling after
Table 2. Behavioral and Survey Results From the Baseline
Analyses
Psychological measure Statistical comparison
Behavioral tasks
Go/no-go r(86) = –.18, p = .09
Stop-signal reaction time r(82) = .25, p = .02
Stroop interference r(86) = .06, p = .58
Visual search r(86) = –.07, p = .52
General surveys
Attention-deficit/hyperactivity disorder r(84) = .21, p < .05
Autism r(75) = .01, p = .93
Maximization r(84) = .17, p = .12
Video-game expertise r(86) = .10, p = .35
Impulsivity
Attentional r(82) = .24, p = .03
Motor r(83) = .12, p = .27
Nonplanning r(81) = –.03, p = .79
Note: Each measure was compared with the number of civilian
casualties inflicted during the video game played in the baseline
session (pretraining session for the two training groups).
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

Reducing Civilian Casualties 1173
*
8
9
10
11
12
13
14
Response Inhibition
Training
Visual Search Training
Civilian Casualties
Before Training
After Training
*
200
210
220
230
240
250
260
Response Inhibition
Training
Visual Search Training
Intended Targets Hit
* *
*
ab
Fig. 4. Mean number of (a) civilian casualties and (b) intended targets hit as a function of training condition and time of test. Asterisks
indicate a significant difference between testing occasions or training conditions (*p < .05). Error bars show standard errors of the mean.
020406080 100
Change in Civilian Casualties
ADHD Score
–8
–6
–4
–2
0
2
4
6
8
10
12
–8
–6
–4
–2
0
2
4
6
8
10
12
020406080 100
Change in Civilian Casualties
ADHD Score
Response Inhibition Training
r(26) = .40
p = .03
Visual Search Training
r(27) = –.08
p = .68
ab
Fig. 5. Scatter plots (with best-fitting regression lines) showing the change in the number of civilian casualties (pretraining minus posttrain-
ing) as a function of the number of self-reported attention-deficit/hyperactivity disorder (ADHD) symptoms, separately for the (a) response-
inhibition training group and (b) visual search training group. On the y-axes, higher values equal lower civilian casualties.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

1174 Biggs et al.
inhibitory training (Verbruggen, Adams, & Chambers,
2012). The current findings suggest another potential
area to which response-inhibition training could be
applied—shooting a firearm. Namely, squeezing the trig-
ger or not squeezing the trigger is akin to a go/no-go
task, albeit one with more complicated lead-up processes
than in standard laboratory tasks. Cognitive training can
improve response-inhibition abilities, which could like-
wise potentially reduce shooting errors due to response-
inhibition failures. The present study provides initial
insight into this relationship by comparing performance
on laboratory-based, response-inhibition tasks with a
simulation designed to mimic the basic response proce-
dure of shooting a firearm. This link will need to be fur-
ther supported by additional evidence, although the
present study provides enticing preliminary results.
The current study also adds to a growing literature
linking gun presence and gun use to cognitive abilities.
For example, previous research has revealed a weapon-
focus effect—individuals remember fewer details about
the perpetrator of a crime if the perpetrator was armed
than if the perpetrator was unarmed (Loftus, Loftus, &
Messo, 1987; for a recent review, see Fawcett, Russell,
Peace, & Christie, 2013). The weapon-focus effect involves
situations in which someone else is holding a weapon,
though recent evidence has demonstrated that wielding a
gun also affects cognition. For example, a person holding
a gun was more biased to see a gun in the hands of oth-
ers (Witt & Brockmole, 2012), and wielding a gun altered
where an individual looked in a scene (Biggs, Brockmole,
& Witt, 2013). The present study extends the previous
research by providing preliminary evidence to link shoot-
ing performance and cognitive abilities, and, more impor-
tant, supports the possibility that shooting performance
could be improved through cognitive training. Here, we
focused on civilian casualties and response inhibition,
but there are many possible action-cognition links
involved in shooting to explore in future research.
From a practical perspective, the current findings sug-
gest potential promise for improving shooting perfor-
mance for a wide range of individuals, including military
and law-enforcement personnel. Similar training efforts
could yield further targeted training regimens so that the
most effective training can be implemented for any given
scenario. For example, competitive sports shooters might
want to enhance their ability to pick up potential targets
through visual search, whereas soldiers might want to
enhance their ability to avoid hitting unintended targets.
Finally, given that this project serves as an initial step
in relating shooting abilities to cognition, there are sev-
eral limitations. First, the current study demonstrated a
significant training benefit, but it remains unclear which
aspect of the training was the primary influence. The
benefit could be due to the stop-signal training task, the
interactive go/no-go training task, or a combination of
the two. The critical point is that training occurred, but
future work will be needed to elucidate the primary
mechanisms. Second, the current design used two active
training groups (rather than an active training and control
group), which limited placebo effects. However, it is pos-
sible that participants expected certain benefits from the
training (e.g., Boot, Blakely, & Simons, 2011; but see
Green etal., 2014), and these expectations could, in the-
ory, have influenced their performance.
It is also important to highlight that the shooting simu-
lation implemented here may or may not adequately
compare with performance among the associated profes-
sional populations (e.g., military and law-enforcement
personnel). Several aspects of the shooting scenarios
were specifically chosen for implementation with an
untrained population (e.g., a targeting reticle helped the
novice participants aim), but additional scenarios are
required to better assess performance among trained
populations. Additionally, the present study isolated a
particular, simulated situation in which civilian casualties
might be inflicted. Participants knew precisely which tar-
gets were hostile and which targets were not—yet civil-
ian casualties were still inflicted. There are numerous
other situations, depending on the rules of engagement,
that could lead to civilians being hit with weapons fire.
Inhibitory control may or may not be important for reduc-
ing civilian casualties across all circumstances. The pres-
ent study demonstrates an initial, basic research
instantiation of a potential link between inhibitory con-
trol and simulated civilian casualties, but future work is
needed to expand this result before it is proper to make
policy suggestions.
In conclusion, the present study represents an impor-
tant step forward in the exciting prospect of better
understanding both shooting and cognition by studying
them in unison. The current evidence was obtained
using a simulated environment to link a particular shoot-
ing error—civilian casualties—to a particular cognitive
ability— response inhibition. This link demonstrates the
potential benefit of predicting shooting performance
through cognitive abilities, but it also highlights the
opportunity to potentially improve shooting performance
through cognitive-training methods. Future work can
examine additional stops along the shooting-cognition
continuum (e.g., target identification and object recogni-
tion) to offer additional insight into shooting performance,
cognitive processes, and the link between the two.
Author Contributions
A. T. Biggs, M. S. Cain, and S. R. Mitroff designed the study.
A.T. Biggs collected and analyzed the data. A. T. Biggs, M. S.
Cain, and S. R. Mitroff discussed the results and helped prepare
the manuscript.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

Reducing Civilian Casualties 1175
Acknowledgments
The authors would like to thank Corporal Ian Seely (U.S. Marine
Corps), Sergeant Anthony Allen (U.S. Army), and Specialist
Sterling Williams (U.S. Army) for volunteering their time to dis-
cuss firearms operations and procedures.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with
respect to their authorship or the publication of this article.
Funding
This research was funded by the Army Research Office under
Contract No. W911NF-13-1-0480.
References
Appelbaum, L. G., Cain, M. S., Darling, E. F., & Mitroff, S.R.
(2013). Action video game playing is associated with
improved visual sensitivity, but not alterations in visual
sensory memory. Attention, Perception, & Psychophysics,
75, 1161–1167.
Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., &
Clubley, E. (2001). The Autism-Spectrum Quotient (AQ):
Evidence from Asperger syndrome/high-functioning
autism, males and females, scientists and mathematicians.
Journal of Autism and Developmental Disorders, 31, 5–17.
Barratt, E. S. (1959). Anxiety and impulsiveness related to psy-
chomotor efficiency. Perceptual & Motor Skills, 9, 191–198.
Berkman, E. T., Kahn, L. E., & Merchant, J. S. (2014). Training-
induced changes in inhibitory control network activity. The
Journal of Neuroscience, 34, 149–157.
Biggs, A. T., Brockmole, J. R., & Witt, J. K. (2013). Armed and
attentive: Holding a weapon can bias attentional priorities
in scene viewing. Attention, Perception, & Psychophysics,
75, 1715–1724.
Biggs, A. T., Cain, M. S., Clark, K., Darling, E. F., & Mitroff,
S. R. (2013). Assessing visual search performance differ-
ences between Transportation Security Administration offi-
cers and nonprofessional searchers. Visual Cognition, 21,
330–352.
Biggs, A. T., & Mitroff, S. R. (2014). Different predictors of
multiple-target search accuracy between nonprofessional
and professional visual searchers. Quarterly Journal of
Experimental Psychology, 67, 1335–1348.
Bissett, P. G., & Logan, G. D. (2011). Balancing cognitive
demands: Control adjustments in the stop-signal paradigm.
Journal of Experimental Psychology: Learning, Memory,
and Cognition, 37, 392–404.
Boot, W. R., Blakely, D. P., & Simons, D. J. (2011). Do action
video games improve perception and cognition? Frontiers
in Psychology, 2, Article 226. Retrieved from http://journal
.frontiersin.org/article/10.3389/fpsyg.2011.00226/full
Boot, W. R., Simons, D. J., Stothart, C., & Stutts, C. (2013).
The pervasive problem with placebos in psychology:
Why active control groups are not sufficient to rule out
placebo effects. Perspectives on Psychological Science, 8,
445–454.
Brainard, D. H. (1997). The Psychophysics Toolbox. Spatial
Vision, 10, 433–436.
Chikazoe, J., Jimura, K., Hirose, S., Yamashita, K. I., Miyashita, Y.,
& Konishi, S. (2009). Preparation to inhibit a response com-
plements response inhibition during performance of a stop-
signal task. The Journal of Neuroscience, 29, 15870–15877.
Eagle, D. M., Baunez, C., Hutcheson, D. M., Lehmann, O., Shah,
A. P., & Robbins, T. W. (2008). Stop-signal reaction-time
task performance: Role of prefrontal cortex and subtha-
lamic nucleus. Cerebral Cortex, 18, 178–188.
Fawcett, J. M., Russell, E. J., Peace, K. A., & Christie, J. (2013).
Of guns and geese: A meta-analytic review of the ‘weapon
focus’ literature. Psychology, Crime & Law, 19, 35–66.
Green, C., Strobach, T., & Schubert, T. (2014). On method-
ological standards in training and transfer experiments.
Psychological Research, 78, 756–772. doi:10.1007/s00426-
013-0535-3
Guerrieri, R., Nederkoorn, C., & Jansen, A. (2012). Disinhibition
is easier learned than inhibition. The effects of (dis)inhibi-
tion training on food intake. Appetite, 59, 96–99.
Houben, K., Havermans, R. C., Nederkoorn, C., & Jansen,
A. (2012). Beer à no-go: Learning to stop responding to
alcohol cues reduces alcohol intake via reduced affective
associations rather than increased response inhibition.
Addiction, 107, 1280–1287.
Houben, K., Nederkoorn, C., Wiers, R. W., & Jansen, A. (2011).
Resisting temptation: Decreasing alcohol-related affect and
drinking behavior by training response inhibition. Drug
and Alcohol Dependence, 116, 132–136.
Jasper, L., & Goldberg, I. (1993). Jasper/Goldberg Adult ADD
Questionnaire. Retrieved from http://healthyparent.com/
ADD-AdultQ.pdf
Kahl, C. H. (2007). In the crossfire or the crosshairs? Norms,
civilian casualties, and U.S. conduct in Iraq. International
Security, 32, 7–46.
Kleiner, M., Brainard, D. H., & Pelli, D. G. (2007). What’s
new in Psychtoolbox-3? Perception, 36(ECVP Abstract
Supplement), 14.
Leotti, L. A., & Wager, T. D. (2010). Motivational influences
on response inhibition measures. Journal of Experimental
Psychology: Human Perception and Performance, 36,
430–447.
Loftus, E. F., Loftus, G. R., & Messo, J. (1987). Some facts about
“weapon focus.” Law and Human Behavior, 11, 55–62.
Logan, G. D. (1994). On the ability to inhibit thought and
action: A users’ guide to the stop signal paradigm. In D.
Dagenbach & T. H. Carr (Eds.), Inhibitory processes in
attention, memory, and language (pp. 189–239). San
Diego, CA: Academic Press.
Logan, G. D., Schachar, R. J., & Tannock, R. (1997). Impulsivity
and inhibitory control. Psychological Science, 8, 60–64.
Macnamara, B. N., Hambrick, D. Z., & Oswald, F. L. (2014).
Deliberate practice and performance in music, games,
sports, education, and professions: A meta-analysis.
Psychological Science, 25, 1608–1618.
Manuel, A. L., Bernasconi, F., & Spierer, L. (2013). Plastic modi-
fications within inhibitory control networks induced by
practicing a stop-signal task: An electrical neuroimaging
study. Cortex, 49, 1141–1147.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from

1176 Biggs et al.
Mastiff. (2013). Reload: Target down [Video game for Nintendo
Wii]. San Francisco, CA: Author.
Menon, V., Adleman, N. E., White, C. D., Glover, G. H., & Reiss,
A. L. (2001). Error-related brain activation during a Go/
NoGo response inhibition task. Human Brain Mapping,
12, 131–143.
Miranda, A. T., & Palmer, E. M. (2014). Intrinsic motivation
and attentional capture from gamelike features in a visual
search task. Behavior Research Methods, 46, 159–172.
Patton, J. H., Stanford, M. S., & Barratt, E. S. (1995). Factor struc-
ture of the Barratt Impulsiveness Scale. Journal of Clinical
Psychology, 6, 768–774.
Pelli, D. G. (1997). The VideoToolbox software for visual psy-
chophysics: Transforming numbers into movies. Spatial
Vision, 10, 437–442.
Schwartz, B., Ward, A., Monterosso, J., Lyubomirsky, S., White,
K., & Lehman, D. R. (2002). Maximizing versus satisficing:
Happiness is a matter of choice. Journal of Personality and
Social Psychology, 83, 1178–1197.
Spierer, L., Chavan, C. F., & Manuel, A. L. (2013). Training-induced
behavioral and brain plasticity in inhibitory control. Frontiers
in Human Neuroscience, 7, Article 427. Retrieved from http://
journal.frontiersin.org/article/10.3389/fnhum.2013.00427/full
Stothart, C. R., Simons, D. J., Boot, W. R., & Kramer, A. F.
(2014). Is the effect of aerobic exercise on cognition a pla-
cebo effect? PLoS ONE, 9, Article e109557.
Thorell, L. B., Lindqvist, S., Bergman, S., Bohlin, G., &
Klingberg, T. (2009). Training and transfer effects of
executive functions in preschool children. Developmental
Science, 12, 106–113.
Verbruggen, F., Adams, R., & Chambers, C. D. (2012). Proactive
motor control reduces monetary risk taking in gambling.
Psychological Science, 23, 805–815.
Verbruggen, F., Aron, A. R., Stevens, M. A., & Chambers,
C. D. (2010). Theta burst stimulation dissociates atten-
tion and action updating in human inferior frontal cortex.
Proceedings of the National Academy of Sciences, USA, 107,
13966–13971.
Verbruggen, F., Chambers, C. D., & Logan, G. D. (2013).
Fictitious inhibitory differences: How skewness and slowing
distort the estimation of stopping latencies. Psychological
Science, 24, 352–362.
Verbruggen, F., & Logan, G. D. (2008). Response inhibition in
the stop-signal paradigm. Trends in Cognitive Sciences, 12,
418–424.
Webb, C. M., & Hewett, K. J. (2010). An analysis of U.S. Army
fratricide incidents during the Global War on Terror (11
September 2001 to 31 March 2008). Fort Rucker, AL: Army
Aeromedical Research Lab.
Witt, J. K., & Brockmole, J. R. (2012). Action alters object iden-
tification: Wielding a gun increases the bias to see guns.
Journal of Experimental Psychology: Human Perception
and Performance, 38, 1159–1167.
Wright, R. G. (2003). Combating civilian casualties: Rules and
balancing in the developing law of war. Wake Forest Law
Review, 38, 129–172.
by Stephen Mitroff on August 20, 2015pss.sagepub.comDownloaded from