Article Text

Download PDFPDF
Cognitive deterioration associated with an expedition in an extreme desert environment
  1. P Maruff1,
  2. P Snyder2,
  3. M McStephen1,
  4. A Collie1,
  5. D Darby1
  1. 1CogState Ltd, Melbourne, Victoria, Australia
  2. 2University of Connecticut, Storrs, CT, USA
  1. Correspondence to:
 Dr Maruff
 CogState Ltd, Melbourne, Victoria 3000, Australia; pmaruff{at}


Background: Prolonged exposure to extreme environments may result in cognitive changes that may influence decision making ability and increasing risk of injury or death.

Objective: To measure the cognitive performance of a healthy man as he completed a 17 day desert expedition.

Method: A computer based cognitive test battery, subjective cognitive rating scale, and measures of physical characteristics were used. Objective cognitive performance was compared with the performance of eight age matched men who remained in their own homes.

Results: The speed of psychomotor, attentional, and executive functions decreased as the expedition progressed, but the accuracy of performance remained unaffected. Although some impairments were large, they resolved completely once the expedition was completed. Subjective ratings indicated that the subject had insight into his failing cognitive performance during the expedition.

Conclusions: Cognitive performance can be measured repeatedly throughout an expedition in an extreme environment. Cognitive impairment may occur.

  • cognitive deterioration
  • desert expedition
  • subjective performance rating

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Anecdotal reports indicate that ongoing exposure to extreme environmental conditions can decrease decision making ability.1,2 Formal scientific studies have corroborated these reports and show that cognitive impairment can occur in extreme environments such as high altitude, extreme cold or heat, and low ocean depths.3,4,5,6,7,8,9,10 In these studies, impairment in attention, psychomotor function, vigilance, and decision making was shown by comparing subjects in extreme environments with people in normal environments or with their own baseline data obtained before the expedition began. These studies also suggest that, although the cognitive impairments can be of moderate severity, the subjects may not have accurate insight into their presence. This is not surprising as subjective impressions of cognitive abilities rarely correlate with objective cognitive deficits in subjects with or without true cognitive impairment.11 However, lack of insight into poor decision making may increase the risk of injury or death in extreme environments.

In general, studies designed to understand the effects of extreme conditions on cognitive function have been limited by the technology used to measure cognitive function—for example, many tests of cognitive function require that a skilled examiner administer them individually. Therefore such cognitive tests must be given remotely—that is, by telephone or radio—or by examiners who are also in the extreme environment so that they too are at risk of experiencing the same cognitive deficits.4,9 Alternatively, cognitive measures have been administered immediately before the extreme conditions are faced, and then again immediately upon safe return.10 Although such a design does support conclusions about cognitive change, it is limited, as cognitive function may return to baseline quickly once subjects are removed from the extreme conditions. Such designs also restrict the ability to study the time course of change or to use more powerful prospective experimental designs that increase the sensitivity of methods designed to detect cognitive change.12

Recent improvements in computer and communications technology have allowed the development of a brief cognitive test battery, CogState, which can be given rapidly and repeatedly and can be self administered.13 This test battery has shown sensitivity to cognitive change in subjects who are fatigued from 24 hours of sustained wakefulness, are suffering the effects of a sports concussion, are at high altitude, or are suffering acute stress.14–16 It therefore has potential for measuring cognitive function in extreme environments.

For this study, a volunteer agreed to have his cognitive function measured progressively while on a 17 day expedition across the Simpson Desert in Australia. We compared his objective and subjective cognitive function with that of a control group of eight age and sex matched controls, also assessed over a 17 day period, who remained in their customary (urban) environs carrying out their normal activities of daily living.



A 44 year old man volunteered to be studied as he undertook the first solo crossing of the Simpson Desert on foot via its most difficult route, the Madigan Line. He had trained for the crossing for about 12 months using endurance exercise and strength training. His weight at the beginning of the expedition was 70 kg and his height was 175 cm. At the end of the expedition he weighed 64.05 kg. He had normal vision and no history of medical illness. He was taking no drugs routinely and reported normal physical health.

The control group consisted of eight adult men aged 42–47 (mean 45.2). All were right handed and had normal or correct to normal vision. None were colour blind and all reported no unusual medical history. All subjects rated themselves as healthy and physically fit.


The expedition required the subject to walk for 17 days. Initially he hauled a cart with supplies and equipment weighing about 100 kg. His food was rationed to deliver 4000 calories (16.7 kJ) per day. On day 7 of the expedition, he jettisoned the cart and began carrying an 80 litre backpack as the amount of supplies he required had been reduced. His crossing required him to walk 556.4 km over more than 700 sand dunes, which ranged in height from 4 to 48 m, as well as non-uniform gibber plains and flood plains. During the first week of the expedition, he reported difficulty in staying warm as he confronted 5–15 knot SE winds and daytime maximum temperatures of 14°C. When he reached the interior of the desert, the weather returned to the expected pattern, with temperatures peaking at 42.5°C. During this time, the expedition was virtually locked in the desert with all main entry and exit points closed because of the continuing and substantial rain. The western section of the desert, which had received substantial rain, revealed an abundance of flora and fauna: camels, dingoes, kangaroos, bearded dragons, and a variety of other lizards, wedge tailed eagles, kites, owls, falcons, galahs, finch, and budgerigars. The interior of the desert saw the flora and fauna quickly disappear and replaced by endless baron dunes, covered in large clumps of spinefex and cane grass. On day 18, the subject returned home where rest, food, and drink were unrestricted.


Physiological and environmental factors

The subject kept a daily diary of the distance travelled, the calories used, and his mean and maximum heart rate. In addition, he rated his level of fatigue and physical condition using a Likert scale (1–10). The anchor points were “This is the least/most fatigued I have ever been” and “This is the best/worst physical condition I have ever been in”. For both scales the “least” anchor corresponded to a rating of 1. Table 1 summarises these data.

Table 1

 Summary of physiological and expedition measures from the 17 day diary

Cognitive function

The cognitive battery CogState was presented on a laptop computer. It consisted of five tasks, which have been described in detail elsewhere.13,14,17 These measures were selected because their performance scores have metric properties optimal for the detection of cognitive change.17 Each task was presented in succession on a green background using playing cards as the stimulus set. On each trial of each task, participants were required to make a yes “d” or no “k” key response. The tasks varied according to what decisions they required the subject to make. These were the detection (psychomotor measuring function), identification (measuring visual attention), monitoring (measuring vigilance), one back (measuring working memory), and learning (measuring associate learning) tasks. Except for the learning task, which required subjects to complete 40 trials, each task required subjects to complete 35 trials.

For all tasks, the speed and accuracy of performance are usually recorded. However, for the analysis of change, we have identified that the log10 transformed reaction times for the detection, identification, monitoring, and one back tasks, and the arcsine transformed proportion correct responses on the one back and learning tasks provide distributions of data with optimal metric properties for the assessment of cognitive change.17

Subjective cognitive function

Immediately after completion of each test session, all subjects were asked to rate how they perceived their own performance across the five measures described above, by answering four questions based on: “That set of tests required you to think quickly. Compared to how you are normally how do you rate this ability right now?” The question was repeated three more times with the terms “remember”, “concentrate”, and “learn” substituted for “think quickly”. For each question, subjects rated themselves on a scale of 0–10, where 0  =  much worse than usual, 5  =  normal, and 10  =  much better than usual. The score on the four items was summed to provide a total subjective impairment rating.


The subject carried on his trek a Zynx ruggerised computer with an external keyboard and traditional computer mouse. It was powered by a 12 V gel cell battery and 10 W solar panel. To become familiar with its requirements, the subject completed the Cogstate battery three times before he began his expedition. Once the expedition began, he completed it on the evening (about 1700 hours) of the first day and then on the evening of every second day until the end of the journey. The entire test battery took about 11 minutes to complete. On completion of the tests, the subject connected the computer (via a data cable) to a Motorolla 9505 satellite phone which was on a satellite (Iridium) network, to send the test results to the offices of Cogstate Ltd in Melbourne, Australia. After completing the Cogstate battery and sending the data, the subject answered the four cognitive subjective performance questions. He then recorded the physical and environmental conditions of the day (table 1).

The eight controls completed the CogState battery three times in one day to become familiar with the test requirements. They then completed it on the evening of the first day and then on the evening of every second day for 17 days. The software was installed on their home computers, and all completed the assessment in their homes. The data from each assessment were sent to the investigators by email.

Data analysis

Data from the practice assessments were not analysed. Data analysis proceeded in three stages. Firstly, the extent to which change in cognitive function and self perceived cognitive ratings of the control group occurred over time was analysed with a series of repeated measures analyses of variance (table 2). Secondly, an estimate of variability in performance over time for the five different CogState performance measures was derived from each analysis of variance by identifying the residual error variance in the model and then computing the square root of this value—that is, within subjects standard deviation (WSD).18 For each outcome measure, the WSD was then expressed as a function of the grand mean (mean of the means for each assessment) to provide an estimate of the coefficient of variation for each measure—that is, coefficient of variation  =  WSD/grand mean. Thirdly, for the expeditionary subject, reliable change indices were computed for each performance measure at each assessment.19 Reliable change indices were computed by subtracting the performance at each assessment from the mean of the performance data collected on the first (day 1) and last (day 21) day of the study. Each difference score was then divided by the WSD for that performance measure to yield a standardised change score. Standardised change scores greater than 1.96 or less than –1.96—that is, p<0.05 two tailed—were classified as abnormal.

Table 2

 Group mean (SD) of performance on the CogState performance measures in the control subjects assessed repeatedly over 19 days in their homes


Table 2 shows the performances of the control group. There were no significant changes in performance over time for any of the measures. Table 2 shows the WSDs for each of these measures and a coefficient of variation. Analysis of variance indicated no significant effect of time for any of the CogState measures. The WSD for each measure was also low, yielding coefficients of variation that were less than 3% for the speed measures and less than 10% for the accuracy measures.

The data for each test for the subject were standardised using WSDs from table 2. Figure 1A,B shows the change from baseline in each of these performance measures. A range of normal performance was established from the WSD as +/−1.96 SD. Performance measures outside this range on any assessment were classified as abnormal.

Figure 1A shows that the accuracy of performance was not affected during the expedition. Figure 1B shows that identification and working memory tasks became abnormal by day 7 of the expedition and continued to decline until day 17. However, cognitive performance had normalised once the expedition was completed and the subject had returned home. Figure 1C shows subjective rating of cognitive function. The subject rated his cognitive function as having declined at day 7 and as continuing to decline at day 9. Lower levels of subjective cognitive performance continued until day 16 and then began to return to normal levels once he had returned home. They then remained at normal levels until the study was complete (day 22).

Figure 1

 Standardised change in performance (reliable change indices) for the subject in each of the CogState accuracy (A), speed (B), and subjective cognitive (C) impairment questionnaire measures over the period of the expedition. Dotted lines indicate normal range for performance change computed from control data (mean +/−1.96SD). B1 indicates the assessments used in the baseline score for computation of reliable change indices. The expedition was completed on day 17, and assessments thereafter were completed in the subject’s home.


The results of this study indicate that there was a significant and substantial deterioration in cognitive performance during the expedition, which normalised on return to a normal environment. Impairment was greatest for tasks that measured the speed of psychomotor and attentional function. The speed and accuracy of working memory and the accuracy of memory functions remained within normal limits. In contrast, the age matched controls who remained in their normal environments performing their normal activities showed no changes in cognitive function over the same period of time. Taken together, these data suggest that the performance change in the subject detected using the CogState battery reflected true cognitive impairment and therefore that some aspect of the expedition led to the transient deterioration in cognitive function. The deterioration was restricted to the speed of decision making (fig 1A), with accuracy not being affected (fig 1B). Importantly, had cognitive function been assessed only at the beginning and end of the expedition, no cognitive changes would have been detected.

The subjective ratings indicate that the subject perceived that his cognitive function was worse than normal throughout most of the expedition (fig 1C). Furthermore, covariation between the objective and subjective cognitive measures is apparent from fig 1A–C. This indicates that the subject did have accurate insight into the fluctuating level of his cognitive function. The experimental design was such that he completed his subjective performance questions immediately after assessing his cognitive function in the evening. In fact, the subjective questions required him to consider the tests that he had just performed and to compare them with how he thought he would perform them under normal circumstances. The CogState tests provide an auditory signal to subjects (a “buzz” sound) whenever errors are made, making it possible for the subject to infer his true level of performance while completing each test. However, the cognitive change observed occurred only for the speed of performance, whereas accuracy was not affected. Hence, the number of error buzzes that he heard during each assessment would have been the same across assessments. Taken together, the experimental data suggest that the subject was aware that his cognitive performance was worse than normal in the middle of the expedition and that this perception was accurate.

What is already known on this topic

  • There is anecdotal evidence that long term exposure to extreme environments can result in reduced capacity for decision making

What this study adds

  • The measurement of cognitive function during exposure to an extreme desert environment can be accomplished

  • Substantial impairment in the speed of decision making was detected which resolved on return to the normal environment

Although the results indicate that the speed of cognitive performance underwent substantial but transient deterioration during the expedition and that the subject did have insight into this deterioration, we cannot determine the specific cause(s) of the impaired performance. Cognitive impairment has also been observed in other studies of chronic exposure to extreme environments,7,8,15 and the process that gives rise to such impairment must reflect the factors—for example, cold, altitude—or combination of factors specific to that environment. The nature and time course of the impairment detected provides some clues as to what may have occurred. For example, the cognitive deterioration occurred gradually. Secondly, the impairment was acute in that it resolved immediately on return to a normal environment. Thirdly, although the cognitive impairment was large, it was confined mainly to performance speed. This pattern of results suggests some physiological or psychological process that was cumulative in its deleterious effects but which ceased when the subject was removed from the extreme environment. The cognitive dysfunction detected is unlikely to reflect an acute confusional state. With the latter, accuracy of performance is impaired and there is general loss of insight into the impairments.20 Furthermore, in our study, the subject showed no difficulty in completing the complex tasks necessary for survival and was also able to dispatch the CogState data by phone and computer.

It is possible that the cognitive deterioration reflected fatigue resulting from sleep restriction, poor quality sleep, or physical exhaustion.21 Although the diaries (summarised in table 1) indicate that the subject slept well and that his physical preparation had equipped him adequately for the expedition (, he did report moderate levels of fatigue and physical exhaustion, which began to increase in the middle of the expedition and continued to increase until he left the extreme environment. It is also evident that, on days 1–5, he expended more than 4000 calories a day. Therefore his cumulative energy debt increased across the early part of the expedition. Although his energy intake became greater than expenditure on days 6–16, his cumulative calorie balance did not become positive until day 13. Thus for most of the expedition, he was expending more energy than he could replace with his rations. Accordingly, he lost 5.5 kg (about 8% of his body weight) over the 17 days of the expedition. The cognitive impairments detected, as well as the moderate levels of fatigue, may therefore have arisen from this energy imbalance. An acidotic-ketotic state resulting from carbohydrate depletion is known to give rise to transient cognitive impairment and fatigue, which resolves once glucose intake is increased.22 This hypothesis would be relatively simple to test by urinanalysis for ketone bodies, but unfortunately urine samples were not collected. We propose to investigate this in future studies.

In summary, this study indicates that it is possible to measure cognitive function throughout expeditions in extreme environments. This measurement indicates that expeditioners may be at risk of cognitive impairment, although the reasons for this are not clear. Finally, the results are limited because the deleterious effects of the extreme environment were detected in a single subject. However, these data now provide the methodological basis for understanding the effects of other extreme environments in groups or individuals.



  • Competing interests: none declared