Protan Response Times to Red Lights in a Mildly Hypoxic Environment

Transcription

1 RESEARCH ARTICLE Protan Response Times to Red Lights in a Mildly Hypoxic Environment H OVIS JK, M ILBURN NJ, N ESTHUS TE. Protan response times to red sighting distance could be reduced by a factor of 2.5 for lights in a mildly hypoxic environment. Aviat Space Environ Med deep red signal lights relative to individuals with normal 2014; 85: Purpose: This study was conducted to determine whether protans color vision (NCV) and deutans ( 6, 15 ). have slower reaction times to red lights than individuals with normal In addition to reduced sighting distances, protans can color vision and to identify whether protan reaction times increase differentially in a mildly hypoxic environment. Methods: Simple reaction also have slower reaction times to red lights ( 5 ). At this point, we should distinguish between choice reaction times (SRT) to a red light-emitting diode (LED) display were measured using the Psychomotor Vigilance Task (PVT) at ground (1293 ft/394 m), times and simple reaction times. Choice reaction time is simulated 12,400-ft (3780-m) altitude, and 20 min after returning to the time required to identify the color of a light when ground. Subjects were 13 individuals with normal color vision (NCV), different colors are possible, whereas simple reaction 12 with a deutan color vision defect, and 4 with a protan color vision time is the time required to respond to the onset of the defect. Results: The mean reaction times increased by 8% with altitude and decreased after returning to ground for all groups. However, the reaction times of the protans were often faster than the NCV mean and cause research has shown that individuals with a deu- appearance of a light. This distinction is important be- never below the NCV 10 th percentile. The only significant difference tan defect can have longer choice reaction times for red between color vision groups was the slowest mean reaction time of the lights than protans when they have to distinguish between green, yellow, and red signal lights ( 1, 3 ). For sim- NCV group was slower than both the pooled dichromats and pooled anomalous trichromats across all conditions by 23%. The number of lapses did not vary with altitude, but the dichromatic subjects had significantly fewer lapses than the trichromatic subjects across all condi- slower (e.g., 550 ms for protanopes vs. 343 ms for NCV) ple reaction times, however, protans can be 25 60% tions. Conclusion: Although protans may be slower to respond to some to react to the onset of a red signal light or red lightemitting diode (LED) even when the intensity is ad- red warning lights, this decrement in performance could not be demonstrated under the conditions of our experiment. Furthermore, the protan group s simple reaction times were not differentially affected by mild justed to be optimum by reaching an asymptote value hypoxia. These results suggest that the red LEDs were sufficiently bright ( 5, 18 ). The increased reaction times may be even longer for these protan observers. for red lights that are dimmer than the optimum intensity ( 5, 18 ). Interestingly, Pun et al. ( 18 ) also reported that Keywords: reaction times, red lights, hypoxia IP:, color vision defects On: Mon,, 01 Oct :31:46 supra-threshold colored signals. Aerospace Medical Association Delivered by deuteranopes Ingenta had slow reaction times to the red LEDs. T HERE ARE TWO types of congenital red-green color vision defects. One type is labeled deutan and the other protan. Individuals with a deutan defect either are missing their middle-wavelength-sensitive cone (M-cone) or have an anomalous photopigment present in their M-cone that has spectral absorption properties similar to their longwavelength-sensitivity cone (L-cone) photopigment to varying degrees. Individuals with a protan defect are missing either their L-cone or they have an anomalous photopigment present in their L-cone that has spectral absorption properties similar to their M-cone photopigment to varying degrees. Those individuals who are missing a cone are referred to as dichromats; those with anomalous photopigments are referred to as anomalous trichromats ( 10 ). Both the deutans and protans can have difficulty in identifying colors used in navigation displays; however, protans also have more difficulty in detecting red lights because they have a significant reduction in brightness sensitivity to long wavelength light ( 12, 20 ). The reduction in brightness sensitivity can have a significant effect on their detection of red signal lights, such that their Jeffery K. Hovis, Nelda J. Milburn, and Thomas E. Nesthus Their reaction times were very similar to the protanopic values. Red warning lights are common in aviation and responding quickly to the warning signal may be critical to safe aircraft operation. The increase in reaction times and shorter sighting distances for red signal lights is one of the primary reasons for the Commission Internationale de l Eclairage (CIE) to recommend that protans should be excluded from holding a commercial pilot s license or certified as an air traffic controller ( 7 ). In addition to the basic reduction in brightness sensitivity to red lights, there is also the concern that protan reaction From the Civil Aerospace Medical Institute, Federal Aviation Administration, Oklahoma City, OK, and the School of Optometry and Vision Science, University of Waterloo, Waterloo, Ontario, Canada. This manuscript was received for review in May It was accepted for publication in July Address correspondence and reprint requests to: Jeffrey K. Hovis, O.D., Ph.D., School of Optometry and Vision Science, University of Waterloo, Waterloo, ON N2L 3G1, Canada; jhovis@uwaterloo.ca. Reprint & Copyright by the Aerospace Medical Association, Alexandria, VA. DOI: /ASEM Aviation, Space, and Environmental Medicine x Vol. 85, No. 11 x November 2014

2 times may be differentially affected by a mild hypoxic environment, such as experienced in general aviation and commercial aircraft. Assuming that the effect of hypoxia is equivalent to lowering the brightness of the light ( 8 ), then the reaction times for protans may increase to a greater degree with decreasing brightness of the red light relative to changes shown for NCV ( 5, 18 ). Given the importance of red warning lights in aviation, we were particularly interested in determining whether protans exhibit longer reaction times to a red light and whether their reaction times are affected differentially by mild hypoxia. METHODS The results presented in this paper are from a study on mild hypoxia and color vision conducted at the Civil Aerospace Medical Institute s hypobaric research chamber located in Oklahoma City, OK. The subject selection and general procedures are described in detail elsewhere, so only a summary will be presented here ( 14 ). Subjects Subjects were recruited through advertisements in Oklahoma City, OK, area newspapers and on flyers placed at area colleges and universities, and were paid by a subject contractor. Each individual was first screened to ensure that his corrected or uncorrected visual acuity Equipment was at least 20/30 in each eye for both near and far Cognitive performance measures included the Psychomotor vision. With one exception, all subjects met the 20/30 Vigilance Task device (PVT-192, Ambu latory acuity criterion. The one exception was a subject with a Monitoring, Inc., Ardsley, NY) (PVT) and eight selected severe deuteranomalous deficiency who also had severe tests from the Automated Neuro psychological Assessment amblyopia in his left eye. Because his results for the binocularly Metric (ANAM) (Automated Neuro psychological As- viewed color vision tests were well within the sessment Metrics, Version 4.0, 2007; Norman, OK). Only range of the other severe deuteranomalous subjects, his the PVT and simple reaction time (SRT) results from the data are included in this study. Subjects requiring correction were asked to wear their untinted ANAM battery are reported in this paper. Preliminary spectacles Aerospace Medical results of Association the other cognitive tests have been presented rather than their contact lenses during screening Delivered and by elsewhere Ingenta ( 16, 17 ). study sessions. Subjects who met the vision requirements completed a medical history form required by the study s medical monitor to make a general assessment of the subjects health. Those cleared by the medical monitor visited a designated Federal Aviation Administration (FAA) aviation medical examiner to complete the medical assessment, which was essentially equivalent to a FAA Class III medical examination with the exception of their color vision status. Disqualifying medical conditions were: deep venous thrombosis; cardio-pulmonary disease; surgery involving the head, thorax, or abdomen within the past 6 mo; use of certain drugs such as b-blockers, sildenafil, and tadalafil; prior diagnosis of decompression sickness; seizures induced by optical stigma due to flickering lights, moving lights, or flashing displays; or diabetes. All subjects were men and nonsmokers between the ages of 18 and 57 yr. Color vision classification was determined by the Nagel Type 1 Anomaloscope using the procedure described for the neutral adaptation state ( 19 ). There were 13 NCV subjects, 8 deuteranomalous, 4 deuteranopes, 2 protanomalous, and 2 protanopes who participated. The mean age of the color-normal group was 35.2 yr (SD ) and the mean age of the colordefective group was 38.8 yr (SD ). The mean ages between groups were not significantly different [ t (27) 5 0.8; P ]. The number of individuals is slightly different from the previous report due to the availability of the different tests. Although the number of protans in the pool is relatively small, the frequency of protan subjects is close to the expected frequency of 2% from the Caucasian color-defective population ( 4 ). The percentage of deuteranopic subjects in our sample was higher than the general population by a factor of 2.0 and the frequency of the deuteranomalous individuals was lower by a factor of Prior approval for all procedures and use of human subjects was obtained from the Institutional Review Board of the Civil Aerospace Medical Institute in accordance with the rights and protections under Title 45, Public Welfare, Department of Health and Human Services, part 46, Protection of Human Subjects ( cfr46.html ). Informed consent was obtained prior to participation, and each subject was free to withdraw from the experiment without prejudice at any time. The PVT is an electronic, hand-held, computerized test-presentation and data capture system for simple visual reaction time. The stimulus is a multidigit display created by small red LEDs presented in a dark background. This display serves as both a stimulus and millisecond reaction time display. The individual LEDs forming a digit were approximately 1 mm by 1 mm separated by approximately 0.3 mm black space. The angular height of the digits at a 40-cm viewing distance was 0.78 degree and the angular size of the stroke width was approximately 0.15 degree. The angular size of the digits corresponds to a 20/180 letter using the equivalent Snellen scale for near visual acuity. Subjects were instructed to press and release a button on the device using their dominant thumb as quickly as possible as soon as the numbers appeared in the display. The task was set for 5-min duration with an interstimulus-interval of 3000 to 7000 ms. Five different PVT units were used in the study. It was not possible to measure the luminance and chromaticity coordinates of the PVT display because the individual LEDs were too small to measure with our available meters. Nevertheless, imaging the Aviation, Space, and Environmental Medicine x Vol. 85, No. 11 x November

3 light from the display of one unit through a monochromator showed that the wavelengths emitted by the LEDs ranged from 630 to 690 nm with a maximum intensity near 660 nm. To estimate the brightness, the fovea detection threshold for the display was measured using a series of neutral density filters before one observer s eye after dark adapting for 5 min. The luminance of the display was 2.7 log units above threshold. Based on Graham and Barlett s threshold data for red lights ( 11 ), this result suggests that the luminance of the red digits was approximately 5.0 cd m 2 2. The ANAM SRT test measured reaction time by presenting the user with a series of white colored asterisk symbols (*) on a blue background. The user was instructed to respond as quickly as possible by clicking the left mouse button with their preferred index finger each time the asterisk appeared. The asterisk was 1.7 by 2.2 cm presented in a 14.1-in screen on an IBM w Think- Pad, Model T41. At a 50-cm viewing distance, the angular subtense was approximately 2.3. The luminance of the asterisk was 104 cd m 2 2 and the luminance of the blue background was 8 cd m 2 2. This test was performed twice in a session; once as the very first test in the ANAM sequence and second as the last test of the ANAM sequence. The two ANAM SRT tests were separated by approximately 35 min. The 5-min PVT was performed shortly after completion of the ANAM tests. The total time required to perform all of the ANAM and PVT tasks was approximately 43 min. Procedure Subjects had four practice trials in the hypobaric chamber at ground with both the PVT and ANAM tests 1 to 3 d before they were scheduled for the altitude trial. On the day of the altitude session, the ANAM test battery and the PVT task were first conducted reach near-asymptomatic values after a session of 5 min. at Aerospace 1293 ft They have been shown to be the more sensitive parameters Ingenta for differentiating between alert and nonalert Medical Association (394 m; Ground); second, after reducing ambient Delivered baro- by sub- metric pressure in the chamber environment to simulate an altitude of 12,400 ft (3780 m, Altitude); then, third, after returning to 1293 ft (394 m) during the postaltitude trial (Post). The 12,400-ft (3780-m) altitude condition was selected based on current regulations in the United States whereby pilots are allowed to fly at a maximum altitude of 12,500 ft (3800 m) without supplemental oxygen. At altitude, subjects rotated to the various color vision testing stations within the chamber every 45 min in a predetermined order so that the sequence of tests was balanced across subjects. Some of these results have been presented elsewhere ( 14 ). After all subjects had completed each of the color vision tasks (after approximately 3.5 h), SRTs were measured as part of the cognitive test battery. The cognitive tests were repeated one final time approximately 20 min after returning the chamber to ground level. The altitude sessions lasted a total of 4.5 h with a short lunch break at the 1.75-h point of the study. Subjects stayed within the chamber during the break and the FAA provided snacks and lunch. The meals and snacks were intentionally light and designed to avoid caffeine and sugary items. Heart rates (bpm) and percent blood oxygen saturation levels (S a O 2 ) were monitored throughout the altitude trials using a multiparameter pulse oximeter (Nellcor Pulse Oximeter, Model 200; forehead electrode placement; Covidien-Nellcor, Boulder, CO). The data were stored on a PC computer using Labview software (Ver. 8.6, National Instruments, Austin, TX). A description of results of the physiological measures associated with the ANAM and PVT testing were presented elsewhere ( 16, 17 ). Statistical Analysis SPSS vers. 20 (IBM, Somers, NY) was used for the statistical analysis. The primary analysis on the response times was ANOVA for repeated measures. Linear regressions were also performed to examine the relationship between the response times and other parameters measured in the study. Statistical significance was based on an alpha level 0.05, unless stated otherwise. RESULTS The PVT provides performance measures for approximately 30 parameters. Because we were interested primarily in the reaction times, this paper will focus on the mean reciprocal reaction time (MRRT, i.e., speed of response), mean of the slowest 10% of the reciprocal reaction times (SRRT), mean of the fastest 10% of the reciprocal reaction times (FRRT), and the number of lapses (i.e., response times. 500 ms). The number of trials used to calculate these values within a session varied from 42 to 54 with 47.3 as the average number of trials. These parameters are often used to measure alertness and jects who have been sleep deprived ( 2 ). We mention this to illustrate the sensitivity of the metric and the potential to differentiate performances between our color vision and altitude conditions. We also examined the pre- and post-pvt fatigue ratings (10 point analog scale) as a subjective measure of fatigue. Fig. 1 shows a combination of box and scatter plots of the MRRT results. The relatively small number of protans makes valid statistical comparisons with other subject groups challenging and so we chose to examine visually where the protan values fell with respect to the normal and deutan range of MRRTs. The dichromatic results are represented by different symbols from their respective anomalous trichromats in the figure. This presentation was selected to determine whether our results were consistent with previous reports that anomalous trichromats, particularly the protanomalous individuals, had better performance than the corresponding dichromats (e.g., protanopes) who have more severe defect ( 6, 18 ). The main result shown in Fig. 1 is that none of the individual protan results fell below the slowest NCV 10 th 1080 Aviation, Space, and Environmental Medicine x Vol. 85, No. 11 x November 2014

4 Fig. 1. Box and scatter plots of the mean reciprocal reaction time (i.e., response speed). Higher values indicate faster reactions. The box represents the distribution of the normal color vision responses between their 25 th and 75 th percentiles, the error bars show the normal color vision 10 th and 90 th percentile values, the solid line is the normal color vision median value, and the dashed line is their mean. The arrows point to the mean reaction speeds of the protan group. percentile responses and only twice did a protan result fall below the 25 th percentile speed. These two instances involved the same 29-yr-old protanomalous subject during the altitude trial at 12,400 ft (3780 m) and, again, after returning to ground level. The deuteranope and deuteranomalous results showed a much wider range of values relative to the protans and NCV groups shown in Fig. 1, but there did not appear to be an obvious systematic difference between their results and those of the other subjects. Although there was no systematic difference in the protan results, Fig. 1 does show the expected results of a decrease in the mean reaction speed at the 12,400-ft value of 1.0 corresponds to zero lapses using this transformation. (3780-m) altitude and a slight improvement after Aerospace returning to ground for all subjects. Because of the small Delivered num- by tan Ingenta was never greater than the NCV 75 th percentile Medical Association The number of lapses for an individual pro- score ber of protan subjects and reports that dichromats (both protanopes and deuteranopes) generally have slower reaction times ( 18 ), statistical analyses were carried out using NCV, pooled anomalous trichromats, and pooled dichromats as between subject factors. The mean values for various reaction speeds are shown in Table I. The ANOVA performed on the MRRT data did, indeed, show that the effect of altitude was statistically significant [F(2,25) ; P, ], but neither the interaction term [F(4,52) ; P ] nor the color vision factor [F(2,26) ; P ] was statistically significant. Although the MRRTs were not statistically different between groups, the mean MRRT values for anomalous trichromats and dichromats were 5 6% faster than the NCV group. The FRRT results in Fig. 2 show a similar trend as with the MRRT. The one difference was that the 29-yrold protanomalous subject was always slower than the NCV 25 th percentile and closer to the NCV 10 th percentile speed. Statistical analysis revealed a significant altitude effect [F(2,25) ; P ], but neither the interaction term [F(4,52) ; P ] nor the color vision factor [F(2,26) ; P ] was statistically significant. As with the previous parameters, Fig. 3 shows that there was a similar altitude trend for the SRRT results and that only once did a protanope have results that were below the 25 th percentile value of the NCV. The more common result was that both the protans and deutans had SRRT response speeds above the mean of the NCV group. The mean SRRT values for both the anomalous trichromats and dichromats were 23% faster than the NCV mean. The ANOVA revealed a significant color vision effect [F(2,26) ; P ], along with a significant altitude effect [F(2,25) ; P ], but the interaction term was not statistically significant [F(4,52) ; P ]. Fig. 4 shows the square root transformation for the number of lapses. This transformation was carried out to create a normal distribution of the results. Note that a of the NCV. Furthermore, the average of the anomalous trichromatic and dichromatic color-defective data was lower than the mean for the NCV group. The ANOVA did not show a significant effect of altitude [F(2,25) ; P ], nor did the interaction term [F(4,52) ; TABLE I. MEAN REACTION SPEEDS AND LAPSES FOR THE NCV, ANOMALOUS TRICHROMATS, AND DICHROMATS. Color Vision Altitude Mean Reciprocal Reaction Time Fastest 10% Reciprocal Reaction Time Slowest 10% Reciprocal Reaction Time Number of Lapses (square root transformation) ANAM Reciprocal Reaction Times Normal ( N 5 13) Initial Ground 4.31 (0.13) 5.50 (0.13) 2.49 (0.16) 2.40 (0.28) 3.77 (0.093) 3780 m 4.01 (0.14) 5.40 (0.17) 2.11 (0.19) 2.64 (0.36) 3.59 (0.12) Post Ground 4.17 (0.14) 5.58 (0.16) 2.23 (0.17) 2.46 (0.32) 3.63 (0.10) Anomalous Trichromatic ( N 5 10) Initial Ground 4.47 (0.15) 5.49 (0.15) 2.99 (0.18) 1.57 (0.21) 3.66 (0.10) 3780 m 4.25 (0.16) 5.39 (0.19) 2.69 (0.22) 1.71 (0.41) 3.49 (0.13) Post Ground 4.35 (0.16) 5.51 (0.18) 2.86 (0.19) 1.56 (0.36) 3.65 (0.11) Dichromatic ( N 5 6) Initial Ground 4.57 (0.20) 5.64 (0.20) 2.95 (0.24) 1.83 (0.41) 3.91 (0.13) 3780 m 4.17 (0.21) 5.31 (0.25) 2.45 (0.29) 2.14 (0.52) 3.96 (0.17) Post Ground 4.43 (0.21) 5.58 (0.23) 2.77 (0.25) 2.41 (0.46) 4.06 (0.14) Standard error values are listed in the parentheses m 5 ; 12,400 ft. Aviation, Space, and Environmental Medicine x Vol. 85, No. 11 x November

5 Fig. 2. Box and scatter plots of the mean of the fastest 10% of reciprocal reaction times (response speed). Higher values indicate faster reactions. The box represents the distribution of the normal color vision speeds between their 25 th and 75 th percentiles, the error bars indicate the normal color vision 10 th and 90 th percentile range, squares are the normal color vision outliers, the solid line is the normal color vision median value, and the dashed line is their mean. The arrows are the protan mean reaction speeds. P ], but the lower number of lapses for the color defective groups was statistically significant [F(2,26) ; P ]. Unexpectedly, the PVT reaction speeds for the colordefective subjects, and the protans specifically, were generally faster than the NCV reaction times. To examine this result further, we examined the simple reaction times measured by the ANAM SRT task. Because the ANAM uses a white asterisk target on a blue background, the luminance contrast of the target would be Fig. 3. Box and scatter plots of the mean of the slowest 10% of reciprocal reaction times (response speed). Higher values indicate faster reactions. The box represents the distribution of the normal color vision speeds between their 25 th and 75 th percentiles, the error bars are the normal color vision 10 th and 90 th percentile scores, square symbols are the normal color vision outliers, the solid line is the normal color vision median value, and the dashed line is their mean. The arrows point to the protan mean reaction speeds. similar across subjects. Fig. 5 shows the average for the reciprocals of reaction times from the first and second ANAM trials at each altitude condition. Note that the individual dichromatic results were above the NCV mean for 12,400 ft (3780 m) and return to ground, whereas the anomalous trichromatic results were more variable, but their individual reaction speeds were generally faster than the 10 th percentile value of the NCV group. The notable exceptions were the two protanomalous individuals at the initial ground trial, who were slightly below the NCV 10 th percentile, and the one deuteranomalous subject with the extremely slow average reaction speed at altitude. We cannot explain this last unusual result. The ANOVA did not reveal a significant altitude [F(2,25) ; P ] or interaction effect [F(4,52) ; P ]; however, color vision classification was significant [F(2,26) ; P ], with the dichromatic reaction times faster than the NCV by about 10% and the anomalous trichromats slower than the NCV group by approximately 2%. These results suggest that the dichromats had faster reaction times in general, whereas the NCV and anomalous trichromats had similar reaction times for the ANAM target. We also examined the results from the five different PVT units to determine if any variation across units was a confounding factor. The number of NCV and colordefective subjects viewing each unit was balanced across units. The ANOVA showed that the mean reciprocal reaction times were not statistically different [F(4,24) ; P ], nor was there a statistically significant interaction with altitude [F(8,46) ; P ] or color vision status [F(2,23) ; P ], so we do not believe that any variation across the units was a factor in our results. The degree of fatigue (both pre- and post-pvt test) increased at 12,400 ft (3780 m) and then decreased after Aerospace Medical returning Association to ground (although still higher than the initial Ingenta ground rating), as expected. The ANOVA Delivered by showed that this effect was significant [F(2,25) ; P ], but there was no significant interaction term between the fatigue ratings and other parameters, including the color vision group [F(4,52) ; P was the largest F value of the interaction terms]. The change in heart rate from ground to altitude and the S a O 2 values at altitude were analyzed to determine whether there were any differences between color vision groups. One-way ANOVA did not reveal any significant differences between the color vision groups for either heart rate [F(2,25) ; P ] or S a O 2 [F(2,25) ; P ] changes with altitude. We also examined correlations of the PVT MRRTs, FRRTs, and SRRTs with subject S a O 2 levels, the change in the average heart rate, and the PVT fatigue rating at 12,400 ft (3780 m). Linear regressions were performed with the subjects pooled into one group because there were no significant differences between color vision groups in these performance measures. None of the linear regressions involving the MRRT and FRRT values reached statistical significance (largest r value was 0.21, P ). Similarly, there was no significant correlation 1082 Aviation, Space, and Environmental Medicine x Vol. 85, No. 11 x November 2014

6 were within the range of the NCV group and the mean reaction speeds for all subject groups showed a similar small decrease (i.e., longer reaction times) with altitude and a subsequent increase in reaction speed after returning to ground level. Essentially, there was no differential effect of mild hypoxia associated with color vision defects with the PVT measures. Although the number of protans was small, none performed below the slower NCV group s 10th percentile value on any of the parameters. Furthermore, there was no systematic trend for the protan results falling below the NCV average. With the exception of the fastest 10% reciprocal response times (FRRT), the mean protan group values on the PVT were usually better than the NCV group mean values. Even though the protan group mean FRRT values were lower than the NVC group means during the 12,400-ft Fig. 4. Box and scatter plots of the number of lapses scaled using (3780-m) trial and the post ground trial, the differences the square root transformation. A value of 1.0 on this scale corresponds were no greater than 3%, which translates into reaction to zero lapses. The box represents the distribution of the normal color vision lapses between their 25 th and 75 th percentiles, the error bars are time differences of only 5.0 ms, hardly an operational the normal color vision 10 th and 90 th percentile scores, squares are the concern. normal color vision outliers, the solid line is the normal color vision Our results differ from previous studies showing that median value, and the dashed line is their mean. The gray dotted line is protans and deuteranopes have slightly slower simple for the pooled anomalous trichromats mean, and the black dotted line is the pooled dichromats mean. The arrows point to the protan mean data. reaction speeds to suprathreshold red lights ( 5, 18 ). The previously reported slower reaction times for deuteranopes may have been confounded by age differences ( 18 ). between the PVT fatigue rating and either S a O 2 level or The age of the subjects in the Pun et al. experiment varied considerably and it was uncertain whether age change in heart rate (largest r ; P ). However, the SRRT values at 12,400 ft (3780 m) were significantly correlated with the post-pvt fatigue rating results were particularly surprising, given that the PVT might have been a confounding factor ( 18 ). Our protan (r ; P for the post-pvt rating), S a O 2 levels displayed a deep red light that had the same wavelength (r ; P ), and change in heart rate (r ; of peak output and luminance as used in Pun et al. s P ). These correlations were not significant for experiment ( 18 ) and the estimated chromaticity coordinates were similar to the red light used by Cole and either ground trial. Brown ( 5 ). Both studies showed a decrease in the protan DISCUSSION response speed at suprathreshold light levels. In addition, others have shown that the wavelength of maxi- The major conclusions from the PVT results were that the protan and deutan reaction speeds to red LED Aerospace lights Medical mum output Association of the PVT was within the spectral region Delivered by where Ingenta the protan brightness sensitivity would be at least Fig. 5. Box and scatter plots of the average of the reciprocals of SRTs from the ANAM first and second trials. The box represents the distribution of the normal color vision responses between their 25 th and 75 th percentiles, the error bars are the normal color vision 10 th and 90 th percentile scores, squares are the normal color vision outliers, the solid line is the normal color vision median value, and the dashed line is their mean. The arrows point to the protan means. 1.0 log unit lower than the sensitivity of color-normals and deutans ( 12, 20 ). There are several possible reasons for the better-thanexpected protan response speeds. The primary reason might be because there were a relatively small number of protan subjects in our sample, so it is possible that these four subjects were not representative of the protan population in terms of their reaction speeds to red lights. This is a valid criticism, but part of the justification for the CIE recommendation that protans should be disqualified from certain aviation professions was that they have slower reaction times to suprathreshold red lights ( 7 ). Our results showed that some protanopes, who should have had the slowest reaction speeds to the red light, actually had better reaction speeds than the NCV group average and that none of the protanope or protanomalous results were below the NCV group s slowest 10 th percentile responses for suprathreshold red light displays. This result weakens the justification for the CIE recommendation. It also raises the issue as to whether a standard can be established on a statistically based policy when there are clearly exceptions to the Aviation, Space, and Environmental Medicine x Vol. 85, No. 11 x November

7 rule. As a group, the protan subjects did not appear to be faster PVT SRRTs and lower number of lapses suggest exceptional (i.e., falling above the fastest 90 th percentile that the color-defective subjects may have experienced for the NCV group) based the ANAM reaction speeds. less fatigue than the NCV group. This result is consistent Their reaction speeds for the ANAM target, where the with the possibility that fatigue may have been a brightness contrast would have been similar across all more important factor in determining reaction times subject groups, were generally similar to the PVT red than brightness sensitivity in this study. display results. The protanopes were generally above The PVT parameters of MRRT, FRRT, SRRT, and number the NCV average speeds and the protanomalous individuals of lapses did prove to be sufficient in measuring re- results were similar to the NCV mean, or below ductions in performance due to mild hypoxia. It was it, so that the average of protan reaction speeds were interesting that SRRT values were significantly correlated very similar to the NCV group across conditions. with the two physiological measures and the fatigue Another possible explanation is that the brightness rating at altitude, whereas the other response speeds sensitivity to red lights for the protan subjects in this were not correlated with the same parameters. This result study was closer to the color-normal value. That is, the suggests that the SRRT may be a better indicator of protan subjects in this study may have had exceptionally a person s inability to adapt to mildly hypoxic environage good sensitivity to red lights compared to the averments, as measured by either fatigue or physiological protan subject. This could be due to their longer changes, than the MRRT and FRRT. wavelength sensitive cones having a peak sensitivity Although the performance of the color-defective subjects, that is shifted to a longer wavelength relative to the particularly the protans, was well within the NCV average protan individual. Although this could be a limits and not differentially affected by hypoxia, it is important possibility, the protan subjects results on the C-100 instrument to remember that our findings are for red lights in this study, which measures relative brightness that are well above threshold for all individuals, regard- sensitivity to orange-red and green LEDs using less of their color vision diagnosis. If the task was deter- flicker photometry, were all below the mean protan setting mining the maximum sighting distance for red signal from a previous study that used a larger group of lights, then the performance of individuals with protan protan subjects ( 13 ). This indicates that the protan subjects deficiencies would be appreciably impaired, because in this study had sensitivities to red lights that were the light intensity would be at a threshold value ( 6, 15 ). no better than the average sensitivity for a larger group There could also be a differential effect in a hypoxic environment of protan subjects. The anomaloscope brightness matches for this viewing situation. for protanopes and color matches for the protanomalous From an operational perspective, our results help define subjects were also typical protan settings. Thus, it is unlikely the parameters for red text warning displays so that that the protans in this study had exceptionally the response of a user with a protan color vision defect good brightness sensitivity to red lights. could have the optimum reaction time. The brightness It is also possible that the subjects in our protan group contrast should be about 3.0 log units above NVC foveal could have had extremely fast reaction times in general. threshold and the angular size of the individual letters or The ANAM results for a target with similar brightness digits should be about 0.75 degree. The 3-log unit contrast contrast across groups showed that the protanopes Aerospace did Medical value is Association very similar to the optimum intensity value reported Ingenta by Cole and Brown for red traffic lights ( 5 have faster reaction times than the mean NCV; however, Delivered by ). the values were never better than the fastest 10 th percentile of the NCV group and the protanomalous subjects generally had ANAM reactions times that were slower than the mean NCV value. As a result, the mean reaction time for the protan subjects, as a group, was similar to the NCV group average values. These results lead us to conclude that, although some protans had good general reaction speeds, it is unlikely that the protan subjects were exceptional in terms of their general reaction speeds and the range of their individual results will likely fall within the range of NCV reaction times for suprathreshold stimuli. Fatigue or sleepiness is often associated with mild hypoxia ( 9 ), and the PVT fatigue ratings did show the expected increase during the altitude trial and subsequent decrease after returning to ground. Although the average fatigue rating was not statistically significant between subject-groups, the NCV group s rating of 6.1 (out of 10) at 12,400 ft (3780 m) was higher than the anomalous trichromats' average rating of 3.9 and the dichromats average rating of 5.1. The lower fatigue ratings for the color-defective groups combined with their ACKNOWLEDGMENTS Research reported in this paper was conducted under the Flight Deck Program Directive/Level of Effort Agreement between the Federal Aviation Administration Headquarters and the Aerospace Human Factors Division (AAM-500) of the Civil Aerospace Medical Institute, and is sponsored by FAA AAM-200 and supported through the FAA NextGen Human Factors Division (ANG-C1). This document is disseminated under the sponsorship of the U.S. Department of Transportation in the interest of information exchange. The United States Government assumes no liability for the contents thereof. We thank Leonard Temme for comments on an earlier version of the manuscript. Authors and affiliations: Jeffrey K. Hovis, O.D., Ph.D., School of Optometry, University of Waterloo, Waterloo, Ontario, Canada, and Nelda J. Milburn, Ph.D., and Thomas E. Nesthus, Ph.D., Human Factors Research Lab, Civil Aerospace Medical Institute, Federal Aviation Administration, Oklahoma City, OK. REFERENCES 1. Atchison DA, Pedersen CA, Dain SJ, Wood JM. Traffic signal color recognition is a problem for both protan and deutan colorvision deficients. Hum Factors 2003 ; 45 : Basner M, Dinges DF. Maximizing sensitivity of the psychomotor vigilance test (PVT) to sleep loss. Sleep 2011 ; 34 : Bergman H, Duijnhouwer F. Recognition of VDU presented colors by color defective observers. In: Proceedings of the Human Factors 1084 Aviation, Space, and Environmental Medicine x Vol. 85, No. 11 x November 2014

8 Society 24th Annual Meeting; October 13 17, 1980; Los Angeles, CA. Santa Monica, CA: Human Factors Society; 1980: Birch J. Worldwide prevalence of red-green color deficiency. J Opt Soc Am A Opt Image Sci Vis 2012 ; 29 : Cole BL, Brown B. Optimum intensity of red road-traffic signal lights for normal and protanopic observers. J Opt Soc Am 1966 ; 56 : Cole BL, Vingrys AJ. Do protanomals have difficulty seeing red lights? In Proceedings of the 20th Session CIE. Vienna: CIE; 1983:1-3. CIE Publication Commission Internationale de L'Eclairage. International recommendations for colour vision requirements for transport. Vienna: CIE; Technical report CIE Connolly DM, Barbur JL. Low contrast acuity at photopic and mesopic luminance under mild hypoxia, normoxia, and hyperoxia. Aviat Space Environ Med 2009 ; 80 : Crow TJ, Kelman GR. Psychological effects of mild acute hypoxia. Br J Anaesth 1973 ; 45 : Deeb SS. Molecular genetics of colour vision deficiencies. Clin Exp Optom 2004 ; 87 : Graham CH, Bartlett NR. The relation of size of stimulus and intensity in the human eye. II. Intensity thresholds for red and violet light. J Exp Psychol 1939 ; 24 : Hsia Y, Graham CH. Spectral luminosity curves for protanopic, deuteranopic, and normal subjects. Proc Natl Acad Sci U S A 1957 ; 43 : Hovis JK. Repeatability of the C-100 colour vision test. Clin Exp Optom 2003 ; 86 : Hovis JK, Milburn NJ, Nesthus TE. Hypoxia, color vision deficiencies, and blood oxygen saturation. J Opt Soc Am A Opt Image Sci Vis 2012 ; 29 :A Kinney JA, Paulson HM, Beare AN. The ability of color defectives to judge signal lights at sea. J Opt Soc Am 1979 ; 69 : Nesthus TE, Dobbins L. Psychomotor performance and physiological effects of hypobaric exposure to 12,400 ft. [Abstract 314]. Aviat Space Environ Med 2008 ; 79 : Nesthus TE, Dobbins L, Milburn N. The effects of mild hypoxia on ANAM test performance for males with normal color vision compared with those with color vision deficiency. [Abstract 166]. Aviat Space Environ Med 2009 ; 80 : Pun HW, Brown M, Lui R. Tinted contact lenses slow reaction time in colour defective observers. Clin Exp Optom 1986 ; 69 : Steen JA, Collins WE, Lewis MF. Utility of several clinical tests of color-defective vision in predicting daytime and nighttime performance with the aviation signal light gun. Aerosp Med 1974 ; 45 : Verriest G. Les courbes spectrales photopiques d'efficacitélumineuse relative dans les déficiences congénitales de la vision des couleurs. [Spectral curves of relative luminous efficiency in congenital deficiencies of color vision.] [Article in French.] Vision Res 1971 ; 11 : Aerospace Medical Association Delivered by Ingenta Aviation, Space, and Environmental Medicine x Vol. 85, No. 11 x November