Cognitive Performance during a 24-Hour Cold Exposure Survival Simulation

Abstract

When a ship grounds in polar regions, evacuees may be required to survive for more than five days before rescue, such that maintaining cognitive performance becomes crucial for survival. Therefore, the purpose of this study was to explore the effects of prolonged cold exposure on cognitive performance. Core temperature (Tc) and cognitive test battery (CTB) performance data were collected from eight participants during 24 hours of cold exposure (7.5°C ambient air temperature). Participants were recruited from those who have regular occupational exposure to cold conditions. Participants were instructed that they could freely engage in minimal exercise that was perceived to aid maintain a tolerable level of thermal comfort. Despite the active engagement, Tc was significantly lower after exposure and the conditions were also sufficient to eliminate the typical 0.5 – 1.0oC circadian rise and drop in core temperature throughout a 24 h cycle. Results showed minimal changes in CTB performance occurred regardless of exposure time. Based on the results and similar findings from previous research, it is recommended that survivors who are waiting for rescue should be encouraged to engage in mild physical activity, which could have the dual benefit of maintaining metabolic heat production and also improving motivation and acting as a distractor from cold discomfort. This recommendation should be taken into consideration during future research and when considering guidelines for mandatory survival equipment regarding cognitive performance.

Full text

Research Article
Cognitive Performance during a 24-Hour Cold Exposure
Survival Simulation
Michael J. Taber,1,2 Geoffrey L. Hartley,1Gregory W. McGarr,1
Dessi Zaharieva,1Fabien A. Basset,3,4 Zach Hynes,3Francois Haman,5,6
Bernard M. Pinet,7Michel B. DuCharme,4,8 and Stephen S. Cheung1,4
1Environmental Ergonomics Laboratory, Department of Kinesiology, Brock University, 500 Glenridge Avenue,
St.Catharines,ON,CanadaL2S3A1
2Falck Safety Services Canada, 20 Orion Court, Dartmouth, NS, Canada B2Y 4W6
3Department of Human Kinetics and Recreation, Memorial University, St. John’s, NL, Canada A1C 5S7
4Marine and Arctic Survival Scientic and Engineering Research Team (MASSERT), Canada
5School of Human Kinetics, University of Ottawa, Ottawa, ON, Canada K1N 6N5
6Nutrition and Metabolism Research Unit, Montfort Hospital Research Institute, Ottawa, ON, Canada
7FacultyofHealthSciences,UniversityofOttawa,Ottawa,ON,CanadaK1N6N5
8Defence R&D Canada, Quebec City, QC, Canada G3J 1X5
Correspondence should be addressed to Michael J. Taber; mt@ca.falcksafety.com
Received  February ; Revised  May ; Accepted  May 
Academic Editor: Edward J. Ryan
Copyright ©  Michael J. Taber et al. is is an open access article distr ibuted under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Survivor of a ship ground in polar regions may have to wait more than ve days before being rescued. erefore, the purpose of
this study was to explore cognitive performance during prolonged cold exposure. Core temperature (𝑇c) and cognitive test battery
(CTB) performance data were collected from eight participants during  hours of cold exposure (.∘Cambientairtemperature).
Participants (recruited from those who have regular occupational exposure to cold) were instructed that they could freely engage
in minimal exercise that was perceived to maintaining a tolerable level of thermal comfort. Despite the active engagement, test
conditions were sucient to signicantly decrease 𝑇caer exposure and to eliminate the typical .–.∘C circadian rise and drop
in core temperature throughout a  h cycle. Results showed minimal changes in CTB performance regardless of exposure time.
Based on the results, it is recommended that survivors who are waiting for rescue should be encouraged to engage in mild physical
activity, which could have the benet of maintaining metabolic heat production, improve motivation, and act as a distractor from
cold discomfort. is recommendation should be taken into consideration during future research and when considering guidelines
for mandatory survival equipment regarding cognitive performance.
1. Introduction
Extreme tourism (i.e., Alaskan and Antarctic cruises) is
becomingmorepopularaslargersectionsofpolaricecap
melt. As a result of expanding marine trac in polar waters,
there is an increase in the possibility of ships grounding on
areas of previously inaccessible shoreline. Evidence of this
possibility can be seen in the number of incidences reported
by the National Transportation Safety Board (NTSB), the
Transportation Safety Board of Canada (TSB), and various
news agencies. Table  shows that since the Majestic Explorer
rammed into a rocky shoal on the Alaskan shoreline in
, leaving one passenger dead and several others injured,
there have been at least  such events in which passengers
may be faced with the possibility of evacuation into harsh
environmental conditions.
It has been reported that a mass rescue operation in sup-
port of an evacuation of a cruise vessel in Arctic waters would
require approximately ve days to complete []. Although
international regulations mandate the amount of food, water,
Hindawi Publishing Corporation
BioMed Research International
Volume 2016, Article ID 8130731, 11 pages
http://dx.doi.org/10.1155/2016/8130731

BioMed Research International
T : Cruise ship emergency events by region.
Name of vessel Year Region Evacuation method
Majestic explorer  Arctic Inatable liferas
Nieuw Amsterdam  Arctic Reoated
Star princess  Arctic Evacuated to another ship
Spirit of   Arctic Inatable liferas
Wilderness explorer  Arctic Reoated
Clipper adventure  Antarctica Freed by Chilean icebreaker
Mona lisa  Arctic Evacuated to another ship
Le conte  Arctic Evacuated to another ship
Wilderness adventurer  Arctic Evacuated to another ship
Clipper odyssey  Arctic Coast Guard assistance
Lyubov orlova  Antarctica Transferred to another ship
Nordkapp  Antarctica Transferred to another ship
MV explorer  Antarctica Lifeboats
Empress of the north  Arctic Coast Guard assistance
Spirit of Alaska  Arctic Coast Guard assist/transferred to another ship
Spirit of glacier bay  Arctic Evacuated to Coast Guard vessel
Ushuaia  Antarctica Evacuated to Chilean nav y vessel
Antarctic dream  Antarctica Free by a research vessel
Ocean nova  Antarctica Evacuated to Argentine Navy vessel
Clipper adventurer  Arctic Coast Guard assistance
Clelia II  Antarctica Assisted by NG Explorer
Polar star  Antarctica Evacuated to Argentine Navy vessel
Sea spirit  Arctic Zodiac capsizes during shore excursion
Silver explorer  Antarctica Damaged by ’ large wave
and equipment International Maritime Organization (IMO)
(): guidelines for ships operating in Arctic ice-covered
waters (I:\CIRC\MSC\-MEPC-Circ), these supplies
are only required to last for three days. Previous research,
conducted over signicantly shorter periods, has shown
that cognitive performance is impaired by environmental
thermal stress []. At present, it is not known whether the
cognitive abilities required to perform vital survival tasks
will be diminished during long-term cold exposure, thereby
reducing the possibility of rescue and ultimately survival
from a polar abandonment from ship or air. More specically,
it is not known whether there are measures that can be
employed by the survivors to mitigate the possible decits in
performance.
erefore, in an eort to address some of these unknown
aspects of long-term survival in cool conditions, this paper
presents cognitive performance ndings from a -hour
experimental cold exposure protocol in which the partici-
pantswereabletoactivelyandvoluntarilycontroltheamount
of physical activity (active engagement) required to maintain
a level of performance perceived to be sucient to complete
a series of cognitive tests. Based on previous research and
duration of the exposure, it was expected that the cool
conditions would impair some aspects of cognition (e.g.,
workingmemoryandexecutivefunctioning)whilehaving
no eect or enhancing other aspects (e.g., reaction time for
simple tasks).
2. Methods
2.1. Participants. e experimental protocol and instrumen-
tation conformed to the standards set by the Declaration of
HelsinkiandwasapprovedbytheResearchEthicsBoardof
Brock University (REB -). Participants were medically
cleared (cardiac stress test) by a primary care physician
for any cardiovascular or neuromuscular symptoms and
provided written informed consent prior to taking part in
the study. An inclusion criterion was regular occupational or
recreational experiences with cold exposure, and participants
came from various professional backgrounds.
2.2. Experimental Design. e overriding design goal was
to have participants thermally stressed to near the limits
of voluntary tolerance, complete with mild hypothermia
and shivering activity, for an entire  h without signicant
participant dropout (see [], for more detail). e experi-
mental design was developed to replicate some of the basic
survival conditions that might be expected following vessel
abandonment in the Arctic. For example, there was limited
access to food, water, mental stimulation, and opportunities
to sleep. To simulate the environmental conditions and
emergency supplies that may be available in a lifeboat or life
ra, no blankets or pillows were provided during the single
continuous session of  hours of cold exposure. Participants
were free to stand or sit in the chamber (described below)

BioMed Research International 
Attention
network
test Groton
maze
learning
test
Groton
maze
learning
test -
recall
Two-back
task
Mental
rotation
task
Cognitive
failure
questionnaire APM
Raven’s
Before cold exposure CTB
approximate time to
complete =80min
CTB - approximate time to complete =40min
F : Full CTB administration protocol.
when they were not taking part in specic testing procedures.
Prior to beginning and also following the -hour cold expo-
sure, participants performed  min of moderate treadmill
walking at % of their maximal aerobic capacity to simulate
the level of physical exertion and activity that might be found
during evacuation and rescue.
To test cognitive performance, a battery of tests that
explored working and long-term memory, vigilance, absent-
mindedness, general mental capabilities, executive function-
ing, information processing ability, and spatial ability in
visual working memory was presented to participants at six
dierent time points throughout the experimental protocol
(baseline,,,,andhoursofexposure,andpostexpo-
sure). Upon arrival on the testing day, participants completed
the cognitive test battery (CTB) as a baseline measure. Aer
exiting the environmental chamber, participants completed a
nal CTB following the postexposure treadmill exercise bout.
e CTB was designed to explore both simple and complex
tasks in an eort to identify where, if any, decits might occur.
ermal comfort and sensation [] were recorded at each
CTB time point.
2.3. Establishment of Ambient Air Temperature Protocol.
Given the limited research associated with long-term expo-
sure, ambient air temperature used for the experimental pro-
tocol was based on pilot testing. Two participants completed
dierent components of the experimental protocol for a
planned  hours, with one additional participant completing
a full  hours of testing. Based on the level of experiences
described by the pilot participants and previous short-term
cold exposure studies [], an ambient air temperature of ∘C
was used for the pilot studies. Body core temperature data
and ratings of thermal comfort were then used to identify the
likelihood that participants would be able to endure/tolerate
theentire-hourexposure.Basedontheresultsofthe
pilottestingandcommentsfromthetwoparticipants,the
experimental test protocol ambient temperature was set at
.  ∘C (see [], for more detail).
2.4. Cognitive Testing. To reduce the likelihood of a learning
eect inuencing the results collected during the actual
administration of the CTB, a familiarization session was
conducted several days prior to the cold exposure. is
familiarization session gave participants the opportunity to
practice sections of the CTB and ask questions about the test
procedures.
As no previous research exploring cognitive performance
during prolonged cold exposure exists, specic cognitive tests
were selected to explore the inuence of exposure on complex
information processing requirements that might be used
duringasurvivalsituation.Toestablishabaselinemeasure
of cognitive self-evaluation and general uid intelligence
(Gf), participants completed a cognitive failure question-
naire (CFQ) [] and Raven’s Advanced Progressive Matrices
(APM) [] ∼ hours before entering the environmental cham-
ber. ese two tests were not completed aer the exposure, as
itisbelievedthatalearningeectmaytakeplaceduringinitial
administration []. For example, Bridger et al. [] reported
a test-retest reliability of . for CFQs completed at  and
 months aer the original administration of the test. Given
the relatively short time period (∼ hours) between the
two test periods, it was expected that the participants would
remember how they had responded on the initial (pretest)
CFQ and Raven’s APM.
To establish a measure of progressive changes throughout
thecoldexposure,participantswerescheduledtocomplete
theattentionnetworktest(ANT)[],Grotonmazelearn-
ing test (GMLT) [], two-back tests (TBT), and mental
rotation tasks (MRT) at predetermined intervals ( h,  h,
 h, and  h) beginning  min aer initial cold exposure
(Figure ). ese measures were also compared against
pre/postexposure scores. e GMLT and TBT are part of the
psychometric measures available from CogState [] and have
been reported to correlate well (𝑟’s = .–.) with similar
psychologicalmeasuresaswellasshowminimalpractice
eects [].
is combination of tests in this integrated CTB has
not been used before for cold exposure research; however,

BioMed Research International
all of the measures have been examined for reliability and
validity in previous cognitive performance studies [, , ].
e following description of each test outlines the areas
of cognitive function believed to be important to decision
making in a survival situation.
Cognitive Failure Questionnaire (CFQ). e CFQ is a -item
questionnaire that represents self-reported cognitive perfor-
mance. e CFQ is measured as a total score ranging from
 to  and is related to four factors of absentmindedness
(memory, distractibility, blunders, and names). In addition,
the CFQ items explore aspects such as spatial orientation
failures, memory lapses, and motor functioning. Items are
scored on a -point Likert scale where  equals “never”
and  equals “very oen.” With high CFQ scores (scored
above ∼), it would be expected that participants may
have considerable diculties completing tasks that require
vigilance (e.g., ANT, Groton maze, and two-back tasks)
in a prolonged cold exposure environment. e CFQ has
been reported to have Cronbach’s alpha value of . and
atest-retestreliabilityof.overa-monthinterval[].
Dependingonthesample,anaverageCFQscoremaybe
between  and  [].
Raven’s Advanced Progressive Matrices (APM).eAPM
is a nonverbal assessment tool designed to measure gen-
eral mental capabilities pertaining to observational skills,
decision-making, problem identication, perception, and
sense making []. A total of  design puzzles with one
piece missing are arranged in an ascending order of diculty
and raw APM scores have been used as an indicator of uid
intelligence [, ]. Individuals are asked to select one of four
possible choices that follow puzzle design pattern rules. For
the purposes of obtaining a baseline measure of cognitive
ability, the APM was completed without a time limit [].
Attention Network Test (ANT). e ANT is a psychometric
tool used to test the eciency of three distinct components
of the human attention network (i.e., alerting, orienting, and
executive control). e test is a combination of a anker task
with arrows and a reaction time task (for a full description
of the test administration methods see [, ]). e ANT
was selected for use during the cold exposure testing as
it has been shown that executive functioning is impaired
when core temperatures are reduced []. Reliability tests for
consecutive ANT performance have shown that a learning
eect exists for the executive functioning, as individuals
progressively get better at ignoring incongruent signals [].
erefore,itwouldbeexpectedthatifthecoldexposure
has no eect on cognitive functioning, reaction times should
improve each time the test is administered. As the ANT was
found to be too onerous during preliminary pilot testing, it
was only administered during the pre-, -hour, -hour, and
posttesting sessions. Ishigami et al. [] suggest that “overall
RT is itself correlated with age (𝑟=.),thenet-work[sic]
scores (𝑟=.and.fortheorientingandexecutivescores,
resp.; ns for the alerting scores), and also with the process
scores (ranging from 𝑟=−. to −.; ns for processes of
Divided attention and Verbal monitor- ing [sic])” (p. ).
Groton Maze Learning Test (GMLT).eGMLTisa
computer-based neuropsychological measure (Cogstate, New
Haven, CT) of working memory functioning (measured by
the maze eciency index) and information processing ability
or executive functioning (measured by the number of errors)
[]. e test consists of a  × grid of square tiles (covering
a hidden pathway –  moves including  turns) presented
to individuals on a touch screen computer surface. When
presented to a participant, the GMLT is randomly selected
from a test bank of  dierent versions of the maze, with
each one equivalent in diculty. Completing one test does
not prepare the individual for subsequent GMLT, therefore,
avoiding a learning eect between tests []. For each CTB
session, the GMLT is presented six times (ve initial repeated
trials and one delayed recall requiring approximately  to
mintocomplete,whichisusedattheendofeachofthe
CTB sessions to test working memory). Trials were timed
(ms) and began automatically when the rst move is made
on the learning trial.
Two - B ack Task (TBT). Similar to the GMLT, the two-back
task (TBT) is a computer-based measure of visual working
memory and attention (CogState, New Haven, CT). e TBT
presents a playing card, shown face up, in the middle of a
screen. Individuals are asked to decide (select “yes” or “no”)
whether a presented card is identical to one shown two cards
before. An interstimulus interval of  seconds is used between
the presentations of  cards. CR selected either the “d”
(no) or “k” (yes) button on a standard QWERTY computer.
Errors on the TBT have been reported to have a signicant
correlation with errors on the GMLT.
Mental Rotation Task (MRT).eMRTconsistedofa
computer-based (available at http://bjornson.inhb.de/?p=)
test of spatial ability in which an individual is given a number
of visual choices that represent a rotated version of a master
image. e master image consisted of small squares arranged
into a pattern presented in an upright position. Each possible
choice contains the same number of small squares; however,
they are arranged in slightly dierent patterns (with the
exception of one correct choice). Diculty in selecting the
correct match to the master image is created by rotating each
of the choices a specic number of degrees to the le or right
(e.g., ∘or −∘). e MRT was selected to explore the
capability to maintain an understanding of spatial orientation
of objects, which was believed to be important during the
nal stages of search and rescue operations (e.g., direction of
aircra in relationship to signaling devices).
2.5. Data Analysis. All data were examined with IBM SPSS
(version ) soware. Prior to performing the statistical
analysis for hypothesis testing, the data were plotted to check
for errors and outliers. Additional checks were performed to
test for the assumptions of normality (Shapiro-Wilkes test)
and homogeneity of variance (Levene’s test). Data collected
from each of the CTB were compared for each of the
dependent variables in a within and between subject design.
Specically, responses for each test were compared across

BioMed Research International 
T : Participant demographic and physiological information.
Participant Measure
Age Height (cm) Weight before (kg) Weight aer (kg) % body fat USG before USG aer ̇
VO2max (mL/min/kg)
  . . . . . . .
   . . . . . 
  . . . . . 
  . . . . . .
  . . . . . . .
  . . . . . .
   . . . . . .
  . . . . . . 
Mean . . . . . . .
(SD) (.) (.) (.) (.) (.) (.) (.)
T : Participant cold exposure experience.
Participant Cold exposure experience
Years of experience Coldest temperature (∘C) Duration of exposure (hours) Last exposure
 −  Within last  months
 −  Within last  months
 −  Within last  months
 −  Within last  months
 −  Within last  months
 −  Within last  months
 −  Within last  months
 Did not respond to questionnaire
Mean . −. 
(SD) . . .
the dierent time points for each participant to identify
possible changes. ese responses were also combined for
all participants to identify general trends in the data based
on the amount of exposure time. Repeated measures analysis
of variance (ANOVA) was used to explore changes in CTB
score for each test. Post hoc analyses were carried out where
appropriate and alpha levels were adjusted according to
Bonferroni corrections.
3. Results
3.1. Initial Cognitive Assessment (CFQ and Raven’s APM). e
results from the CFQ and Raven APM were used as measures
of standardized cognitive processing. Due to known learning
eects, these tests were only performed at baseline and were
notrepeatedwithinthisstudy.emeanCFQscorewas.
(SD = .) (within the normal range for North American
population) and all recorded CFQ scores fell within the %
condence interval []. e untimed administration of the
Raven APM scores ranged from  to  with a mean of .
(SD = .). e scores were found to be normally distributed
and based on standardized norms for a North American
population []; the scores range from the th to >th
percentile.
3.2. Responses to Exposure. Given the intense nature of
the applied research setting, participant sample size varied
throughout the test protocol based on voluntary dropout.
Tables  and  provide participant demographic information
as well as an overview of previous occupational or recre-
ationalexposuretocoldenvironments.Althoughtheexclu-
sion criteria were designed to limit the dropout rate, these
uctuations in sample size reect situation in which it might
beexpectedthatwithinagivenpopulationforcedtoevacuate
a commercial airliner or cruise ship some individuals may not
survive until rescue arrives.
Given that the participants were recruited for their
past experience, Table  details the relevant cold exposure
information. From the table it can be seen that on average
the participants have more than  years of cold exposure
experience in temperatures ranging from −∘Cto−∘C.
SubjectRatingofDiculty. On a subjective rating of dif-
culty where zero represented not dicult at all and 
represented extremely dicult/need to withdraw from the
study, participants rated the experience as an . Seven of
the eight (%) individuals indicated that there was at
least one point throughout the trial that they believed they
would have to voluntarily withdraw from the testing. Two

BioMed Research International
of the participants voluntarily removed themselves from the
experimental protocol—one at .h and the other at  h.
CTB results include the scores of these two participants for
the period in which they remained in the trial (i.e., during the
baseline and rst  hours). Another participant performed
theentirehoursbutrequiredtheuseofthethermalblanket
fromhonwards.Overall,thissuggeststhattheconditions
were suciently taxing physically and mentally even for
this motivated and self-selected participant pool. Seven of
eight participants reported minimal sleep (∼. h) over the
 hours. None of the participants found the bedspace to be
a safe haven or comfortable, and most declined to use the
bedspace or prematurely removed themselves from it over the
course of cold exposure.
Core Temperature (𝑇c). No participants were removed from
the experiment due to core temperature reaching .∘C,
though two participants used the thermal blanket at various
points throughout the testing. For most participants, core
temperature generally decreased ∼.∘C (SD = .) within the
rst  h of exposure and then stabilized at that approximate
level for the remainder of the -hour exposure. Fluctuations
within a range of .∘C occurred during this latter “stable”
period, but overall the participants were able to suciently
thermoregulatethroughshiveringandsomeactiveengage-
ment of mild exercise. Overall, this drop in core temperature
was found to be signicant (𝐹(4,32) =.,𝑝<.),
with a ∼.∘C(SD=.)dropincoretemperaturebetween
thebaseline(pretest)andallothertimepoints(,,,
and  hours of exposure). A similar examination of both
thermal comfort and sensation did not reveal any signicant
dierences across the  hours of exposure. Importantly,
the thermal exposure also eliminated the typical .–.∘C
circadian rise and drop in core temperature throughout a -
hour cycle [], such that the true level of hypothermic strain
exceeded the ∼.∘Cabsolute𝑇cdecrease for much of the
exposure.
ANT Results. e individual mean scores of the three net-
works: alerting, orientation, and conict, were .ms (SD
= .), . ms (SD = .), and . ms (SD = .),
respectively. e mean total reaction time for correct trials
was . ms (SD = .). e test was administered at four
dierent time point: before cold exposure;  hours;  hours,
andaercoldexposure.Aone-wayANOVAindicatedthat
there were no signicant dierences between scores based on
administration time for alerting, orientation, or conict. No
signicant dierences in total mean reaction time were found
for any of the test blocks.
GMLT Results.AspartofthecomputerizedCogStateportion
of the CTB, the GMLT was completed a total of six times
duringthecourseofthisstudy(before,h,h,h,h,
and aer). Given the test protocol, individual GMLT were
presented to the participant seven times (initial test sequence,
ve consecutive presentations, and one recall presentation
aer completing the TBT). Aer the initial presentation, the
nal maze in the initial block (Test Code GMLT-) error rate
was compared with the recall maze (Test Code GMLT-R)
Total errors
Test code
GMLT-5
GMLT-R
36.00
36.50
37.00
37.50
38.00
23456781
Participant
0
2
4
6
8
10
12
14
Tcore (∘C)
F : Total GMLT error based on GMLT- and GMLT-Recall
duringthepostexposuresession.ewhitecirclesrepresentthe
mean 𝑇cof participants at the time of completing the recall GMLT.
using a Wilcoxon Signed Rank Test. No signicant changes
in performance were noted across the trials. However, when
comparing the same two presentations of the Groton maze
for the postexposure session, the results indicate that there
was a signicant dierence between the number of errors
committed (𝑍=−., 𝑝= .). Figure  shows that
almost all of the participants committed more errors while
completing the recall GMLT during the postexposure CTB
session. No other signicant ndings were found.
TBT Results. e two-back task was administered during the
same time points as the GMLT. A repeated measures ANOVA
wasconductedtoexploretheTBTspeed,variability,accuracy,
and number of errors. No signicant ndings were found
regardless of trial administration time.
MRT Results.ementalrotationtask(MRT)measuredthe
accuracy,speed,errors,andtimefortheimageswithineach
test block. Table  displays the mean reaction time (RT) and
accuracy for each of the test blocks.
AFriedmanTestrevealedthattherewasasignicant
reduction in MRT accuracy (𝑛=,𝑝=.).Posthoc
analyses indicated that the signicant dierence occurred
between rst test aer beginning the exposure ( h) and aer
 h of exposure (Figure ). Posttest results indicate that when
completing the computer-based MRT, participants did not
require signicantly more time or commit more errors when
compared to the pretest MRT. Results further indicated that
pretest and posttest computer-based MRT correlate well with
one another (𝑟(6) =.,𝑝= .). No other signicant

BioMed Research International 
T : Mean (SD) reaction times and accuracy for the MRT based
on test blocks.
Test scor e Test block (hour)
Before     Aer
Reaction time
(ms) (SD)
.
(.)
.
(.)
.
(.)
. 
(.)
.
(.)
. 
(.)
Accuracy
(number of
errors)
.
(.)
.
(.)
.
(.)
.
(.)
.
(.)
.
(.)
235781
Participant
Accuracy (number of errors)
Pretest
6hours
12 hours
18 hours
24 hours
Posttest
0
1
2
3
4
F : MRT errors committed by each participant across all
trials.
ndings were found for speed or accuracy regardless of the
time at which the test was completed.
3.3. Combined Cognitive Eects. is section of the results
addresses some of the combined eects of the experimental
conditions on cognitive performance. Correlation analyses
reveal that there were several signicant relationships that
existed within the CTB results; however, no eect of time was
found for performance. Table  displays the correlation table
for all of the tests as they relate to one another.
4. Discussion
isstudyaimedtosimulateaprolongedsurvivalscenario
thatmightoccurfollowingashiporplaneincidentina
cold environment. e primary goal was to elicit a sus-
tained moderate thermal stress throughout  h, including
a decrease in core temperature and elevation in metabolism
through shivering. Additionally, we simulated many of the
attendant situational factors, including isolation, boredom,
food quality and availability, and sleep restriction. Rather
than the restricted movement or voluntary physical activ-
ity in traditional hypothermia research, we permitted self-
engaged physical activity within the connes of the testing
chamber to replicate what might occur in a survival scenario.
Overall, despite these challenging experimental conditions
(conrmed by the participant diculty ratings), cognitive
performance (measured by the CTB) did not signicantly
alter throughout the course of the prolonged  h of cold
exposure compared to baseline values taken before cold expo-
sure. Together, this suggests that cognitive performance may
be maintainable through sustained cold exposure, assuming
that severe hypothermia can be avoided.
Despite previous cold exposure research showing a decre-
ment in simple and choice serial reaction times, memory,
sustained vigilance, and target tracking [–], others have
shownthatlittleornochangesincognitiveperformancewill
occuroverprolongedexposureifindividualsaregiventhe
opportunity to self-regulate the amount of protective clothing
worn or where exercise is used during the exposure sessions.
For example, Slaven and Windle [] showed that there were
no signicant decreases in cognitive performance in serial
RT, Sternberg one-letter or seven-letter recall accuracy or
speed.esendingswerebasedonfourdaysofconsecutive
testing in which the ambient air temperature was ∘C,
.∘C, .∘C, and .∘C (resp.) and participants were able to
select between three options of thermal protection. Similarly,
Banderet et al. [] found that over a ve-day cold exposure
session (ambient air temperatures ranged from −∘Catnight
to −∘C during the day) which included physical activity,
signicant dierences in cognitive performance were only
found for individuals who were hypohydrated at or below
.% of total body weight. In addition to the Slaven and
Windle [] and Banderet et al. [] studies within an
applied setting, Baddeley et al. [] suggest that the lack of
changes in cognitive performance found during exposure to
.∘C water for approximately  minutes was due to highly
motivated divers. Exposure times for all of these studies were
considerably less than those carried out in this examination
of cognitive performance.
Flouris et al. [] show deterioration of vigilance and
reaction time within the rst  minutes of exposure to
−∘C ambient air temperature, while a meta-analysis of
cold exposure studies revealed that cognitive performance is
decreasedbyanaverageof%intemperatureat
∘Corless
[]. Meta-analyses [, ] have identied thermally induced
reductions in cognitive performance that are most oen
observed when tasks are highly complex, require sustained
vigilance, and place a considerable load on working memory.
However, within these studies, the decrements to cognitive
performance have been limited to specic domains such as
working memory and vigilance [], while eects have also
been shown in long-term memory recall [].
e disparity in research ndings on the eects of cold-
induced changes in cognitive performance has previously
been explained by suggesting that the environmental stimuli
(hot or cold ambient temperatures) act as a distractor [,
, ] or as form of arousal [, ]. Based on the results
from each of the cognitive tests, it appears that cognition was

BioMed Research International
T : Correlation table for CTB administration.
Correlations
CFQ APM Alerting Orientation Conict ANT
RT
GMLT
errors
MRT
speed
MRT
accuracy
CFQ score
Pearson correlation 
Sig. (-tailed)
N
APM score
Pearson correlation . 
Sig. (-tailed) .
N
ANT alerting
Pearson correlation −. .∗
Sig. (-tailed) . .
N
ANT orientation
Pearson correlation . −. . 
Sig. (-tailed) . . .
N   
ANT conict
Pearson correlation . . . . 
Sig. (-tailed) . . . .
N    
ANT reaction time
Pearson correlation . . −. .∗∗ .∗∗ 
Sig. (-tailed) . . . . .
N     
GMLT total errors
Pearson correlation . −. .∗. . . 
Sig. (-tailed) . . . . . .
N      
MRT speed
Pearson correlation . . . . . . . 
Sig. (-tailed) . . . . . . .
N       
MRT accuracy
Pearson correlation .∗−. −. . −. −. . .∗∗ 
Sig. (-tailed) . . . . . . . .
N        
∗Correlation is signicant at the . level (-tailed).
∗∗Correlation is signicant at the . level (-tailed).
not signicantly aected when examining overall changes
during the long-term exposure. Given the paucity of ther-
moregulatory research assessing cognitive performance in
long-term (more than  hours) cold exposure, it is interesting
to note that Pilcher et al. [] and Pietrzak et al. [] reported
that high intensity/short-term exposure had a greater neg-
ative inuence on performance than less intense/long-term
exposure during testing. ese ndings have been supported
by research suggesting that cognitive performance above
ambient temperatures of approximately ∘Cwillhavemin-
imal or no changes, whereas ambient conditions below this
temperature result in deleterious eects []. Finally, Færevik
et al. [] reported that minimal physical activity can be
used to minimize a reduction in core temperature over long-
term cold exposure ( hours). Together, this suggests that
there exists a zone of optimal cognitive performance from

BioMed Research International 
maintaining thermoneutrality, similar to that suggested by
Hanin [] and the extended-U hypotheses proposed by
Hancock and Warm []. Both models suggest that there
is a specic zone in which individuals will perform at
maximal levels; however, performance becomes degraded if
individuals are expected to perform tasks outside this optimal
zone. ermal stressors that do not increase the level of
arousal beyond the optimal zone should, therefore, not be
expected to adversely aect the performance of skills that are
well rehearsed or require minimal cognitive eort.
One explanation for the limited changes to cognitive
performance in this study is believed to be the amount of mild
exercise performed by the participants and it was noted that
individuals generally stood throughout the entire experimen-
tal protocol, indicating that it was too cold to sit for any length
of time. It was observed throughout the cold exposure trials
that participants would engage in mild physical activity such
as hopping in one spot, swinging arms around their torso
in a hugging motion, or vigorously rubbing their limbs aer
every test session that required them to sit for any period
of time. Færevik et al. [] showed that minimal exercise in
cold conditions aects core temperature, suggesting that “-
min periods of moderate cycling leg movements every  min
reduced shivering intensity, improved heat balance, slowed
corecooling,andhadapositiveeectonthesubjective
perception of thermal comfort and reduced cold sensation”
(p. ). Specically, in an eort to explore the eects of
exercise on body core while in water, Færevik et al. []
reported that, in −∘Cambientair,
∘Cwaterwith–cm
waves, participants rate of core cooling was signicantly less
when they performed moderate (sustainable) cycling for 
minutes every  minutes. In fact, it was reported that not
only did the moderate exercise decrease the rate of core
cooling, it also signicantly increased heat production to the
point that a % gain was observed [].
A second explanation for the results may be related to
the experimental design of the project and appear to support
previous research suggesting that minimal or no dierences
in cognitive performance should be expected when the
intensity of the stressor (the cold ambient air in this case)
islow[,,,].esendingsalsoappeartosupport
theresultsreportedbySlavenandWindle[]inwhich
they describe that choice RT tests and short-term memory
were unaected aer seven days in a simulated submarine
in distress at .∘C. Similarly, Giesbrecht et al. [] showed
no signicant dierence in the performance of simple tasks
based on cold exposure. Enander [] noted that there were
no decrements in simple RT tasks when participants were
exposed to ∘C ambient air temperature over a period of  to
 minutes (see also []). Participants in this prolonged cold
exposurestudyweregiventheopportunitytoengageinany
activity that would help them stay warm enough to endure
the full -hour exposure protocol. Despite the initial drop in
core temperature, this active engagement potentially provides
a coping mechanism during the testing and possibly could
equate to a survival advantage in abandonment. ere did not
appear one denitive change in cognitive performance over
the course of the experimental session. For example, there
was no dierence in attention related results (ANT and TBT),
while there was a minor (nonsignicant) shi (more errors)
in the working memory aer  hours of exposure, and MRT
results showed that, at  and  hours, there was a tendency
for the participants to require more time to complete the
questions and committed more errors (also not found to be
signicant).
Finally,itcouldbearguedthatanotherexplanationforthe
ndings is related to the level of stimulation present through
the experimental session. e results might suggest that the
level of arousal associated with the cold ambient air, con-
stant shivering, cognitive testing, limited sleep, and conned
conditions fell within an optimal zone of functioning for the
selected group of participants [, ]. e only exception to
this argument of the conditions falling within the optimal
zone was found for the MRT errors. It is possible that the
signicantly higher number of errors in the MRT (when
compared to the performance at  hours of exposure) could
suggest that the level of arousal in the initial part of the testing
was sucient to inuence the performance. As previously
mentioned, the group of participants was specically selected
for this study to ensure a high success rate of completion.
It may be that the intensity of the experimental protocol
was ideally suited to provide a tolerable level in which
performance was not aected []. For example, the ANT
results were consistent with values reported by Weaver et al.
[] who found an overall mean reaction time (RT) of .
(SD = .), an alerting mean score of . (SD = .),
orienting mean score of . (SD = .), and conict mean
score of . (SD = .).
e results indicated that there were signicant corre-
lations between a number of the CTB measures. Given the
specic cognitive tests used in this study and the reported
validity, it would be expected that there would be strong
correlations between and within particular components of
theCTB.Itwas,however,somewhatunexpectedthatno
changes occurred in the latter portion of the prolonged
exposure. Given the fact that no time related correlations
(e.g., negative relationship) were found, it can be assumed
that the participants were suciently stimulated to overcome
the expected inuence on fatigue.
4.1. Main Contributions. As the majority of previous cold
exposure research is conducted over considerably shorter
periods of time (e.g., < hours), the primary contributions
from this study are related to the extension of cognitive per-
formance data over a much longer time frame. Additionally,
the novel experimental protocol, which allowed individuals
to actively engage in mild exercise, situates the data in more
realistic conditions. For example, unless injuries preclude
movement and assuming normothermic core temperatures
as seen in this study, it is unlikely that survivors of a vessel
abandonment will passively sit in one position while they
continue to cool to the point that they no longer have the
capability to help themselves.
4.2. Limitations. With no other changes despite a slight
decrease in core temperature, constant shivering throughout

 BioMed Research International
the exposure period, lack of sleep, and minimal food, it could
be argued that it would be dicult to explain which factor(s)
allowed participants to remain at a nearly constant level of
cognitive performance. e limited number of participants
tested in this study and the changes, both positive and
negative depending on the type of measure, may have been
duetofatigueorexposureoracombinationofseveral
other factors. Additionally, the individual dierences in the
responses to the GMLT may have obscured the eects of
the cold response. Signicant changes may also have been
mediated by increased arousal levels associated with the
distractive nature of the cold exposure [].
4.3. Conclusions. In summary, despite a realistic survival
simulation involving  hours of prolonged cold exposure,
moderate decreases in core temperature, and sustained shiv-
ering, cognitive performance was largely maintained. is
suggeststhat,aslongassignicanthypothermiaisprevented,
survivors may be capable of maintaining a range of simple
through complex cognitive tasks for at least the rst 
hours of abandonment. One potential contributor to this
performance maintenance may be the allowance of mild, self-
engagedphysicalactivity,whichcouldhavethedualbenetof
maintaining core temperature and also improving motivation
and acting as a distractor from the cold discomfort.
Competing Interests
e authors declare that there is no conict of interests
regarding the publication of this paper.
Acknowledgments
e authors are indebted to the participants for their spirit
and enthusiasm throughout the study. e project was funded
by a grant from the National Search & Rescue Secretariat
(SAR New Initiative Fund supported by Transport Canada)
through a contract from Public Works & Government Ser-
vices Canada (Solicitation no. -/A) and a grant
from the Natural Sciences & Engineering Research Council
of Canada (no. -, S. S. Cheung). M. J. Taber was
supported by a Ministry of Research and Innovation Post-
Doctoral Fellowship, and S. S. Cheung was supported by
a Canada Research Chair. C. van Nijen assisted with data
analysis.
References
[] IMO, ReporttotheMaritimeSafetyCommittee,thSession,
Agenda Item , International Maritime Organization (IMO)
Sub-Committee on Radio Communications and Search and
Rescue, .
[] T. M. M¨
akinen, “Human cold exposure, adaptation, and per-
formance in high latitude environments,” American Journal of
Human Biology,vol.,no.,pp.–,.
[] F. Haman, O. L. Mantha, S. S. Cheung et al., “Oxidative fuel
selection and shivering thermogenesis during a - and -
hcold-survivalsimulation,”Journal of Applied Physiology,vol.
, no. , pp. –, .
[]A.P.Gagge,J.A.J.Stolwijk,andJ.D.Hardy,“Comfortand
thermal sensations and associated physiological responses at
various ambient temperatures,” Environmental Research,vol.,
no.,pp.–,.
[] A. L. Vallerand, P. Tikuisis, M. B. Ducharme, and I. Jacobs,
“Is energy substrate mobilization a limiting factor for cold
thermogenesis?” European Journal of Applied Physiology and
Occupational Physiology,vol.,no.,pp.–,.
[] D.E.Broadbent,P.F.Cooper,P.FitzGerald,andK.R.Parkes,
“e Cognitive Failures Questionnaire (CFQ) and its corre-
lates,” British Journal of Clinical Psychology,vol.,no.,pp.–
, .
[] D. A. Bors and T. L. Stokes, “Raven’s advanced progressive
matrices: norms for rst-year university students and the
development of a short form,” Educational and Psychological
Measurement, vol. , no. , pp. –, .
[] D. A. Bors and F. Vigneau, “e eect of practice on Raven’s
Advanced Progressive Matrices,” Learning and Individual Dif-
ferences, vol. , no. , pp. –, .
[]R.S.Bridger,S. ˚
A. K. Johnsen, and K. Brasher, “Psycho-
metric properties of the Cognitive Failures Questionnaire,”
Ergonomics,vol.,no.,pp.–,.
[] J. Fan, B. D. McCandliss, T. Sommer, A. Raz, and M. I.
Posner, “Testing the eciency and independence of attentional
networks,” Journal of Cognitive Neuroscience,vol.,no.,pp.
–, .
[] A. Vom Hofe, G. Mainemarre, and L. C. Vannier, “Sensitivity
to everyday failures and cognitive inhibition: are they related?”
European Review of Applied Psychology,vol.,no.,pp.–,
.
[] CogState, CogState Research, CogState, Melbourne, Australia,
.
[]A.Nyaradi,J.K.Foster,S.Hicklingetal.,“Prospectiveasso-
ciations between dietary patterns and cognitive performance
during adolescence,” Journal of Child Psychology and Psychiatry
and Allied Disciplines,vol.,no.,pp.–,.
[] Y. Ishigami and R. M. Klein, “Repeated measurement of the
components of attention using two versions of the Attention
Network Test (ANT): stability, isolability, robustness, and relia-
bility,” Journal of Neuroscience Methods,vol.,no.,pp.–
, .
[] J. Raven, “e Raven’s progressive matrices: change and stability
over culture and time,” Cognitive Psychology,vol.,no.,pp.–
, .
[] J.C.Wallace,S.J.Kass,andC.J.Stanny,“ecognitivefailures
questionnaire revisited: dimensions and correlates,” Journal of
General Psychology,vol.,no.,pp.–,.
[] N. Goode and J. F. Beckmann, “You need to know: there is a
causal relationship between structural knowledge and control
performance in complex problem solving tasks,” Intelligence,
vol.,no.,pp.–,.
[] J. Raven, J. C. Raven, and J. H. Court, Manual for Raven’s
Advanced Progressive Matrices, Oxford Psychologists Press,
Oxford, UK, .
[] J. W. MacLeod, M. A. Lawrence, M. M. McConnell, G. A. Eskes,
R.M.Klein,andD.I.Shore,“AppraisingtheANT:psychometric
and theoretical considerations of the attention network test,”
Neuropsychology,vol.,no.,pp.–,.

BioMed Research International 
[] M. B. Spitznagel, J. Updegra, K. Pierce et al., “Cognitive
function during acute cold exposure with or without sleep
deprivation lasting  hours,” Av i ati o n , Spa c e, and E nvir onm e n-
tal Medicine,vol.,no.,pp.–,.
[] Y. Ishigami, G. A. Eskes, A. V. Tyndall, R. S. Longman, L.
L. Drogos, and M. J. Poulin, “e Attention Network Test-
Interaction (ANT-I): reliability and validity in healthy older
adults,” Experimental Brain Research,vol.,no.,pp.–
, .
[]R.H.Pietrzak,P.Maru,L.C.Mayes,S.A.Roman,J.A.
Sosa, and P. J. Snyder, “An examination of the construct validity
and factor structure of the Groton Maze Learning Test, a new
measure of spatial working memory, learning eciency, and
error monitoring,” Archives of Clinical Neuropsychology,vol.,
no. , pp. –, .
[] R. S. de Oliveira, B. M. Trezza, A. L. Busse, and W. J. Filho,
“Use of computerized tests to assess the cognitive impact of
interventions in the elderly,” Dementia & Neuropsychologia,vol.
, no. , pp. –, .
[] J. M. Bland and D. G. Altman, “Measuring agreement in method
comparison studies,” Statistical Methods in Medical Research,
vol.,no.,pp.–,.
[] L.C.Lack,M.Gradisar,E.J.W.VanSomeren,H.R.Wright,and
K. Lushington, “e relationship between insomnia and body
temperatures,” Sleep Medicine Reviews,vol.,no.,pp.–
, .
[] S.R.K.Coleshaw,R.N.M.vanSomeren,A.H.Wol,H.M.
Davis,andW.R.Keatinge,“Impairedmemoryregistrationand
speedofreasoningcausedbylowbodytemperature,”Journal
of Applied Physiology: Respiratory Environmental and Exercise
Physiology,vol.,no.,part,pp.–,.
[] A. Enander, “Eects of moderate cold on performance of
psychomotor and cognitive tasks,” Ergonomics,vol.,no.,
pp.–,.
[] A. D. Flouris, D. A. Westwood, and S. S. Cheung, “ermal
balance eects on vigilance during -hour exposures to −∘C,”
Aviation Space and Environmental Medicine,vol.,no.,pp.
–, .
[] G. M. Slaven and C. M. Windle, “Cognitive performance
over  days in a distressed submarine,” Avia tion S pace a nd
Environmental Medicine, vol. , no. , pp. –, .
[] L. E. Banderet, D. MacDougall, D. Roberts, D. Tappan, and M.
Jacey, Eects of Hypohydration or Cold Exposure and Restricted
Fluid Intake Upon Cognitive Performance,DTICDocument,
Natick, Mass, USA, .
[] A. D. Baddeley, W. J. Cuccaro, G. H. Egstrom, G. Weltman, and
M. A. Willis, “Cognitive eciency of divers working in cold
water,” Human Factors,vol.,no.,pp.–,.
[] J. J. Pi lcher, E. Nadler, and C. Busch, “Eects of hot and cold
temperature exposure on performance: a meta-analytic review,”
Ergonomics, vol. , no. , pp. –, .
[] P. A. Hancock, J. M. Ross, and J. L. Szalma, “A meta-analysis
of performance response under thermal stressors,” Human
Factors,vol.,no.,pp.–,.
[] W. H. Teichner, “Assessment of mean body surface tempera-
ture,” Journal of Applied Physiology,vol.,no.,pp.–,
.
[]H.D.Ellis,“eeectsofcoldontheperformanceofserial
choice reaction time and various discrete tasks,” Human Factors,
vol. , no. , pp. –, .
[] Y. L. Hanin, “Emotions and athletic performance: individual
zones of optimal functioning model,” European Year of Sport
Psychology, vol. , pp. –, .
[] P. A. Hancock and J. S. Warm, “A dynamic model of stress and
sustained attention,” Human Factors,vol.,no.,pp.–,
.
[] H. Færevik, R. E. Reinertsen, and G. G. Giesbrecht, “Leg
exercise and core cooling in an insulated immersion suit
under severe environmental conditions,” Aviat i on, S p ace, a nd
Environmental Medicine, vol. , no. , pp. –, .
[] G. G. Giesbrecht, J. L. Arnett, E. Vela, and G. K. Bristow, “Eect
of task complexity on mental performance during immersion
hypothermia,” Aviation Space and Environmental Medicine,vol.
,no.,pp.–,.
[] R. Gillingham, A. A. Keefe, J. Keillor, and P. Tikuisis, “Eect
of caeine on target detection and rie marksmanship,”
Ergonomics,vol.,no.,pp.–,.
[] B. Weaver, M. B´
edard, J. McAulie, and M. Parkkari, “Using the
Attention Network Test to predict driving test scores,” Accident
Analysis and Prevention,vol.,no.,pp.–,.