Introduction

Attention refers to concentrating and choosing a relevant stimulus among other irrelevant information1. In daily life, attention is essential for several activities, such as driving and adequate performance in highly demanded jobs (e.g., radar monitoring and air traffic control)2,3. According to Lezak4, attention is crucial for adequately functioning in other cognitive domains4. Converging lines of evidence suggest that the attention system consists of four subdomains5,6,7,8,9: intrinsic alertness, behavioral inhibition, focused attention, and sustained attention. Intrinsic alertness refers to the internal control of arousal without an external clue, while behavioral inhibition is the ability to control inadequate responses. Focused attention reflects the ability to respond to correct targets, and sustained attention is frequently described as the ability to concentrate over prolonged periods to detect only relevant events.

Attention subdomains can be measured by Continuous Performance Tests (CPT)2. The basic paradigm of a CPT task is the serial presentation of target and non-target stimuli, and the subject's task is to respond to targets and control inadequate responses10. Since CPTs proved a valid evaluation tool, numerous tests have been developed, including the Continuous Visual Attention Test (CVAT)11,12,13. The standard CVAT task consists of stimulus presented at different speeds (1, 2, and 4 s) with two different target probabilities (high and low). The CVAT detects visuomotor reaction time (RT = intrinsic alertness), variability of reaction time (VRT = sustained attention), omission errors (OE = focused-attention), and commission errors (CE = response-inhibition). Although the interpretation of OE, CE and RT are straightforward, the same is not true for VRT14.

The standard CVAT takes 15 min to complete and is validated for clinical use. It has been used in various disorders such attention-deficit hyperactivity disorder (ADHD), fibromyalgia15, obstructive sleep apnea13,16, hepatitis17, and chronic kidney disease18. As current neuropsychology batteries are long and time-consuming, an ultrafast task could be clinically valuable, especially when a quick screening is required. Accordingly, an ultrafast version of the CVAT (90 s) was created and has recently been administered in exploratory studies involving patients with Alzheimer's disease (AD), mild cognitive impairment (MCI), and COVID-1919,20,21. However, further studies are still necessary to confirm the clinical validity of the ultra-fast version as a screening tool.

Besides overall task length, the two versions differ by target probability (TP) (20% and 80% in the 15-min test vs. only 80% in the 90-s test). Another difference between the two tests is that in the 15-min CVAT there are three interstimulus time intervals (ISI) while the 90-s CVAT only includes the highest speed presentation of stimuli. The stimulus duration is the same in the two CVAT’s versions (250 ms). Stimulus-onset asynchrony (SOA) denotes the amount of time between the start of one stimulus, S1, and the start of another stimulus, S2 (SOA = ISI + stimulus duration). SOA in the 90-s CVAT is always 1 s while the 15-min CVAT is equally divided into three different SOAs (1, 2, and 4 s). Current evidence suggests that manipulation in CPT task parameters (TP, SOA, task length) may affect response characteristics, some of which are used as markers of the attention subdomains22,23,24. For instance, when the probabilities or temporal characteristics are varied, the meaning of CPT measures, particularly OE and CE, may change.

The reasons for the choice of the time length (90 s), TP (80%), and SOA (1 s) in the ultra-fast version of the CVAT were based on previous investigations. Firstly, other authors have demonstrated that tasks consisting of 20–60 trials and a minimal duration of 52 s are sufficient to generate reliable VRT and RT metrics25,26,27. Secondly, when no-go trials are infrequent (high TP), responses tend to be quick and less variable, leading participants to readily commit errors on no-go trials. Conversely, when no-go trials (low TP) are frequent, the response rate decreases, and reactions become slower and more variable. Thirdly, a linear relationship between shorter ISI and increased CE has been described but this relationship seems to be moderated by other factors such as TP28,29. Moreover, Steinborn et al.30, have reported that responses are usually slow in trials with a short foreperiod paradigm. An empirical study to verify the adequacy of the choice of these parameters for the ultrafast version has not been investigated yet.

The first objective of this study was to verify the effect of time length, TP, and SOA on the CVAT variables—15-min version (OE, CE, RT, and VRT). Therefore, the purpose of the first part of the experiment was to establish whether high TP and high SOA were the most appropriate conditions for the 90-s task. In the present study we also performed a comparison between the two versions of the tests. For a more systematic examination we used a structured approach as described by Schumann et al.31.

In healthy subjects, the present study aimed to analyze the effect of TP, and time length on the CVAT variables across the 15-min task (1st objective) and verify correlations and agreements between the 15-min and the 90-s CVAT versions on the paired variables (OE, RT, VRT, and CE) of the two CVAT versions (2nd objective).

Methods

Participants

We initially enrolled 228 participants .The exclusion criteria were set prior to data collection and were as follow: age > 60 or < 18 years; taking antipsychotic or anti-epileptic medication at any time and psychotropic drugs that could interfere with attention performance; reduced kidney or hepatic function; past head trauma and loss of consciousness; current alcohol/substance use disorder; pre-existing neurologic or psychiatric disorders; non-corrected hearing or visual impairments; and previous cognitive impairment, and illiteracy. We classified as illiterate the subjects who never learned how to read.

After data collection we removed all subjects who did not engage in the task. Subjects who did not engage in the task commonly exhibited a typical behavior where this is immediately noticeable by intermittent extreme slow and error-prone sequences32,33. In the statistical analysis subsection, we objectively defined how we identified these subjects.

The same subjects performed the two versions of the CVAT (order of the tasks was counterbalanced).

Participation was voluntary, and the local ethical committee approved the research protocol (CAAE:30547720.3.0000.0008). The study was performed by the Helsinki Declaration. Informed written consent was obtained from the participants.

General procedures (attention assessments)

All participants performed the two versions of the CVAT, a Go/No-Go test approved for clinical use (Fig. 1). One version lasts 15 min, and the other 90 s. The same participants of the 15-min task also performed the 90-s task on the same day. For each participant, the order of the two tasks (15-min and 90-s) was counterbalanced. The break between the 2 tasks was of 5 min. The participants received practice trials to both tasks. Failing the practice session consisted of committing one CE or one OE.

Figure 1
figure 1

Schematic overview of the Continuous Visual Attention attentional test (CVAT): (a) the CVAT starts with on-screen instructions: “In this test, the computer alternately displays the indicated figures in the center of the screen. You must press the spacebar using your dominant hand as fast as possible whenever the star appears in the center of the screen. If the other figure appears, you should not press the space”. (b) The target (star) remains on the screen for 250 ms. (c) The non-target (diamond) stays on the screen for 250 ms. Variables provided by the test: average Reaction Time of the correct responses (RT), Intraindividual Variability of Reaction Time (VRT = standard deviation of the RTs during the test), Omission errors (OE), and Commission errors (CE). The CVAT is available for research and clinical use (licensed psychologists) upon request of Prof. Sergio L. Schmidt. The test is available in English, Spanish, and Portuguese.

Full size image

The test sessions were performed in a quiet room, where only the participant and examiner were present during the evaluation. The participants were placed in front of the computer with a distance between the center of the monitor and the eyes of approximately 50 cm. The subjects had visual acuity equal to or better than 20/30 in both eyes, and glasses were used when needed. Before starting the task, the examiner instructed the subject to press the spacebar as fast as possible each time the specific stimulus (target) appeared on the monitor (two geometric figures: star, target, diamond, non-target). Each stimulus was displayed for 250 ms. The test started with instructions and a practice session. The practice sessions took 10 s. A second practice session was administered if the participant failed the first one. Only participants who succeeded in the practice session (first or second) were allowed to continue the experiment. In the present study, all participants succeeded in the practice tasks. The measures included OE, CE, average reaction time of correct responses (RT), and variability of correct reaction times (VRT). VRT was estimated considering a per-person measure of the standard deviation (SD) of individual RTs for the correctly signalled targets. We also estimated the intraindividual variability of the RTs by the coefficient of variability (CV = SD/mean RT)34. This metric (CV) was used to exclude the possible influence of RT on VRT.

15-min CVAT version (Fig. 2)

This version had six blocks, each consisting of three subblocks of 20 trials (Fig. 2). These subblocks had different SOA(s): 1, 2, and 4 s, and different TP: high 80% and low 20%. The first 5 min of the task consisted of the first two blocks; from 5 to 10 min, it included blocks 3 and 4, and from 10 to 15 min, the task included the last two blocks.

Figure 2
figure 2

15-min CVAT: 15-min CVAT consists of three identical tasks with 5 min. The test has 6 blocks with different TP (high 80% and low 20%) and 3-time levels (0–5 min—1 level, 5 -10 min—2 level, 10–15 min—3 level). For each block, the test has 3 subblocks of 20 trials. These subblocks have different SOA: 1, 2, and 4 s. TP target probability, SOA stimulus-onset asynchrony.

Full size image

90-s CVAT version (Fig. 3)

Figure 3
figure 3

90-s CVAT version: the task consists of 90 trials, 72 correct targets, and 18 non-targets. The task only has one TP (80%) and SOA is always equal to 1 s. TP target probability, SOA stimulus-onset asynchrony.

Full size image

In this version (90-s CVAT), the main task consisted of 90 trials (two figures presented, one each time: target or non-target), 72 correct targets, and 18 non-targets. The SOA was always equal to 1 s. The test took 90 s to complete. In summary, the 90-s task is like the 15-min task, except that it included only the higher TP (80%) displayed at the highest speed (SOA = 1 s).

Statistical analysis

The plan for removing subjects after data collection involved excluding participants who exhibited the following: (i) OE ≥ 50%; (ii) CE = 100% without OE (in this case the participants pressed the space bar to all stimuli, irrespective of being a target or a non-target); (iii) an average RT or VRT 3 standard deviations above the average.

We first addressed the distributions of average RTs in both versions of the CVAT. Normality was tested by the Kolmogorov–Smirnov test (K–S) (Fig. 4). Following the recommendations outlined by Schumann et al.31, we investigated the skewness of the data. Skewness was used to measure the extent to which the distribution of values deviates from symmetry around the mean. A value of zero means the distribution is symmetric, while a positive skewness indicates a greater number of smaller values, and a negative value indicates a greater number of larger values. The values for asymmetry between − 2 and + 2 are considered acceptable to prove normal univariate distributions. Previous investigations have reported that it is reasonable to consider a normal univariate normal if skewness is between − 2 to + 235,36. We also measured the inter-correlations of the CVATs variables by using Pearson’s correlation coefficients.

Figure 4
figure 4

Histogram displaying distributions of RTs in the 15-min and 90-s CVAT: (a) and (b) the histogram and K–S show that RTs are normally distributed in both versions of the CVAT task. RT reaction time.

Full size image

First objective

A repeated-measure design was used to verify the effect of TP (low and high), SOA (1,2, and 4 s), and Time (5, 10, and 15 min) on the CVAT variables of the 15-min version of the CVAT. Initially, a repeated-measures MANCOVA was used to test differences in the CVAT variables, considering the two different TPs, the three SOAs, and the 3 Time Levels. The within-subjects factors were Time (0–5, 5–10, 10–15 min; 3 levels), TP (high-80% and low-20%; 2 levels), and SOA (2, 1, and 4 s; three levels). Box’s M test was used to assess the homogeneity of the covariance matrices. Then, respective univariate repeated-measures ANCOVAS were performed using Greenhouse–Geisser correction for sphericity. A P-value of 0.05 (two-tailed) indicated significance. For the MANCOVA and each one of the univariate ANCOVAs, η2 (Eta squared) was computed to calculate the effect size of the results (η2 = 0.01 is considered a ‘small’ effect size, 0.06 a ‘medium’ effect size, and 0.14 a ‘large’ effect size). We used sex and age as covariates.

To avoid harming the statistical power, covariates are typically included in statistical models to account for the variance they explain in the dependent variables. As we have a repeated-measure design, it was also possible to perform the MANOVA and respective ANOVAs without covariates. Therefore, based on the study design (repeated measures) and considering that this approach could increase the statistical power, we performed the MANOVA and respective ANOVAs without covariates. As the results using ANCOVAS or ANOVAS did not show any significant discrepancies, we only reported the resuls derived from the ANCOVAS.

Second objective

A correlation analysis was conducted to verify the relationship between the attentional performance in the two versions of the CVAT. Thus, for each CVAT variable, the Pearson product-moment correlation coefficient was calculated.

Intraclass correlation (ICC) was used to estimate inter-version agreement on raw data derived from the CVAT variables. We used the two-way mixed model because we assumed a random effect of the CVAT data and a fixed effect of the two modalities. Intraclass correlation coefficients were calculated using an absolute agreement definition. To give clinical significance to the ICCs, we considered the F Test with a true value equal to 0.4 instead of 0.

Bland–Altman plots (B–A) were also used to assess agreement between measurements on the same subject. For the B-A analysis, a scatter plot was constructed to plot the difference between the paired measurements on the y-axis and an average of both measurements on the x-axis. A central horizontal line on the plot represented the bias (mean difference in values obtained with both measurements). The standard deviation (SD) of the differences between paired measurements was used to construct horizontal lines above and below the central horizontal line to represent 95% limits of agreement (LOA; mean bias ± 1.96 SD). Conclusions on agreement and interchangeability of both measurements were made based upon the width of these LOA compared to a priori stipulated clinical criteria defined in the power analysis. For each CVAT variable, we also checked if there was heteroscedastic distribution (i.e., whether the magnitude of differences increases proportionally to the size of the measurement).

Clinical significance of the differences

Statistically significant differences between the different parameters of the 15-min CVAT (TP, SOA, and Time) for each CVAT variable (OE, CE, RT, and VRT) were analyzed considering clinically relevant differences instead of relying only on the statistical significance (1st objective). In addition, differences between the two CVAT versions in the averaged paired variables were also interpreted considering the clinical relevance of the differences (2nd objective).

Previous studies on the two versions of the CVAT in different clinical conditions had found the average differences in all CVAT variables that reached clinical relevancy. Based on these studies, differences higher than 35 ms for RT and 15 ms for VRT, a percentage of OE greater than 5% of the total number of correct targets, and a percentage of CE greater than 10% of the total non-targets were found to have sound clinical interpretations13,17,18,20,37.

Results

Demographic data

After applying the exclusion criteria, the final sample consisted of 205 individuals with a mean age of 30.71 ± 10.65 years old (minimum age = 18 years, maximum age = 59). Most of the participants were male (68%). The education level ranged from 2 to 22 years of education. Illiterates were not included.

Frequency distribution of RTs in the two versions of the CVAT

As mentioned in the methods section, we first addressed the distributions of the RTs in the two test versions. Inspecting the histograms and the K-S statistics (Fig. 4) indicated that the RTs were normally distributed in both versions of the CVAT.

Skewness and inter-correlations

Skewness of the data

In the 15-min CVAT we found a highly skewed distribution only for OE variable. In the 90-s task a highly skewed distribution was also observed for OE. All other measures of skewness were between − 2 to + 2.

Correlation matrix of the CVATs variables (Fig. 5)

For the 15-min task, Pearson correlation coefficients indicated that VRT was highly correlated with CV, moderately correlated with RT as well as correlated with OE. RT variable was weakly correlated with CV and negatively correlated with CE. A similar pattern was found for the 90-s task.

Figure 5
figure 5

Correlation Matrix: for the 15-min CVAT, the analyses of the Person correlation coefficients show that VRT is highly correlated with CV, moderately correlated with RT, and correlated with OE. RT variable is weakly correlated with CV and negatively correlated with CE. In the 90-second task a similar pattern is observed. CE commission errors, OE omission errors, RT reaction time, VRT variability of reaction time, CV coefficient of variability, M mean, SD standard deviation.

Full size image

1st objective: effects of Time, TP, and SOA on the CVAT variables (15-min version)

The inspection of the average raw data indicated an absence of clinically significant differences among the VRTs and OEs across the time and in different TPs and SOAs. Conversely, RT and CE were strongly affected by TP. Considering the clinical significance of the differences, Time and SOA did not affect all four CVAT variables (Table 1).

Table 1 Absolute average differences (|Δ|) within the CVAT parameters (time, ISI, and TP) for each CVAT variable (RT, VRT, OE, and CE).
Full size table

The repeated-measures univariate ANCOVAs indicated statistically significant effects of Time (F = 3.728, df = 1.979, P = 0.025, η2 = 0.018), SOA (F = 8.809, df = 1.812, P < 0.01, η2 = 0.042), and TP (F = 29.418, df = 1.00, P < 0.01, η2 = 0.127) on CE. For the RT variable, a similar result was found (Time: F = 7.960, df = 1.806, P = 0.01, η2 = 0.038; SOA: F = 5.773, df = 1.747, P = 0.05, η2 = 0.028; TP: F = 81.315, df = 1.00, P < 0.01, η2 = 0.287). For the OE variable, only TP reached significance (F = 5.759, df = 1.00, P = 0.017, η2 = 0.028). The VRT variable was not affected for TP (F = 0.732, df = 1.00, P = 0.393, η2 = 0.004) SOA (F = 0.668, df = 1.981, P = 0.512, η2 = 0.003), and Time (F = 0.851, df = 1.958, P = 0.426, η2 = 0.004) (Fig. 6).

Figure 6
figure 6

Effect of target probability on the CVAT variables during the 15-min task: (a) VRT per block in the 15-min test: the inspection of the average raw data indicates an absence of clinically significant differences among VRT across time and TP. VTR is not affected by time on task and TP. (b) RT per block in 15-min test: inspection of raw data shows that RT is strongly affected by TP but not by time on task. There are statistically significant effects of TP and time on task on RT. (c) %OE per block in the 15-min test: raw data show no clinically significant differences among OEs across time and in different TP. TP affects OE but not time on task. (d) % CE per block in the 15-min task: data inspection shows that CE is strongly affected by TP but not by time on task. There are statistically significant effects of Time and TP on CE. CE commission errors, OE omission errors, RT reaction time, VRT variability of reaction time, TP target probability.

Full size image

2nd objective (Fig. 7)

Figure 7
figure 7

Correlations and agreements between the 15-min and the 90-s CVAT’s versions. (a) VRT regression between 90 s and 15 min: Person correlation indicates a significant correlation between the two versions in the VRT data. (b) RT regression between 90 s and 15 min: Person correlation shows a significant correlation between the two versions. (c) ICC: the graphic shows a significant interclass correlation(p = 0.048) between the two CVAT versions for the VRT variable. (d) Bland–Altman of VRT: Bland–Altman plot for values of VRT for 15-min and 90-s tests indicates that values are distributed around the mean value for the differences and inside the 95% confidence interval. VRT variability of reaction time, ICC intraclass correlation.

Full size image

As expected, there was a significant correlation between the two versions in the VRT variable (r = 0.54, P < 0.001). Moreover, a significant correlation was also found when intraindividual variability of RTs was estimated by CV (r = 0.45, P < 0.001). Pearson correlation coefficients reached statistical significance in both RT and CE variables. For the OE, the correlation coefficient did not reach significance.

The ICC achieved significance for VRT (ICC = 0.596, df = 204, p = 0.048) and CV (ICC = 0.567, df = 204, p = 0.01). In contrast, the ICC for the other CVAT’s dependent variables did not reach statistical significance. It should be stressed that the null hypothesis was not settled at r = 0. Instead, the null hypothesis considered a correlation of 0.4.

Bland–Altman plot for values of VRT for 15-min and 90-s tests showed that values were distributed around the mean value for the differences and inside the 95% confidence interval. Considering the clinical criteria (Δ = 15 ms), more than 95% of the differences between measurements were not outside the clinically pre-defined relevant limits.

Discussion

Considering the 15-min CVAT, VRT was unaffected by TP, SOA, and time on task. However, RT and CE were strongly affected by TP in the 15-min version of the test. We found a significant interclass correlation between the two CVAT versions (15-min and 90-s for the VRT variable). Moreover, the Bland- Altman plot graphically confirmed agreement for the VRT variable.

Frequency distribution of the RT variable in the two versions of the CVAT

The hypothesis of normality of the RT’s distribution was not rejected in both versions of the CVATs. There is a lot of debate in the neuropsychological field regarding how to correctly calculate intraindividual variability of reaction times. The most common way involves SD of the RTs, CV (SD/average RT), and ex-Gaussian parameters24,38. However, previous investigations have indicated that different metrics generated similar results27,39. Here, we estimated intraindividual variability of reaction times by SD and CV because we did not find evidence that RT distribution was abnormal.

First objective: 15-min task

Based on the clinical significance of the differences, time on task did not affect VRT, OE, CE, and RT. The 15-min CVAT has three identical tasks with 5 min, so time on task was analyzed by comparing the participant's performance in each 5-min task. Our findings indicate that the participants did not experience mental fatigue and that there was no learning effect during the task. In agreement, Bezerra et al., using the same task, found that CVAT variables were not affected by time length in the control group (healthy participants)16.

RT and CE were strongly affected by TP in the 15-min task. The fact that we observed faster RTs and more CE under high TP agrees with several studies involving healthy individuals23,40,41. Here, we propose that a high TP may be interpreted as a more challenging part of the task, and consequently, participants are more engaged, resulting in faster RTs. Indeed, previous investigations have shown that participants self-report being more involved (having more task-related thoughts) on high-response frequency attention tasks than on low-response frequency Tasks40,42.

While participants may be more engaged in high response frequency attention tasks, our results indicate that they also commit more CE. Recent evidence suggests that a high-response condition involves greater pre-motor control and significant prepotent motor activity, which is only observed in high TP conditions (rare no-go tasks)43. This study also found that increased prepotent motor activity in high TP conditions (rare no-go tasks) was followed by greater frontocentral P3-EEG amplitudes, suggesting that inhibitory control is significantly involved in high-frequency attention tasks43. Therefore, we suggest that the participants made more CE in the high TP condition due to a higher inhibitory-control demand in this condition.

Intriguingly, time on task, SOA, and TP did not affect VRT, but average RT was affected by TP. This finding suggests RT does not have a strong effect on VRT. This is the first study showing that sustained attention (VRT) is unaffected by TP, SOA, or time on task in a healthy young population. More importantly, this finding suggests that VRT may be a trait-like characteristic of an individual, as proposed by other authors44,45.

Second objective: 90-s x 15-min task

The high correlation between the two tasks in three CVAT variables indicated that an increase in RT and CE in the 15-min task was associated with an increase in the same variables in the 90-s task. The absence of agreement showed that TP affected these valuables and obscured any possible agreement.

Noteworthy is that we found a high correlation and significant agreement between the two tasks for the VRT variable. Also, ICC achieved significance for VRT but not for the other CVAT variables. It has been proposed that VRT is a trait-like characteristic of an individual44,45; consequently, individuals with a high VRT in one task will also present a high VRT on different tasks, which may explain our results. Of note, VRT is considered an index of central nervous system functioning and is associated with white matter integrity in healthy subjects27,46,47,48.

As previously mentioned, VRT has been frequently used to measure sustained attention, often defined as the ability to concentrate over prolonged periods trying to detect events49,50. Another theory defines sustained attention as maintaining a consistent behavioral response to task-related stimuli during continuous and repetitive activity51. However, the minimum amount of time on a task necessary to characterize sustained attention is not described in both definitions. Here, we showed that sustained attention can be measured by a test with only 90 s of duration with similar results to the 15-min CVAT test, at least in a healthy population.

In terms of clinical application, the 90-s test is helpful in screening for disorders of which VRT is known to be affected, such as ADHD, COVID-19 with gastrointestinal symptoms, MCI, and AD13,19,21,48,52,53. However, in clinical conditions where RT is affected, it is essential to investigate how this variable is affected by TP and SOA to verify if the ultrafast test may be performed instead of the 15-min CVAT. The ultrafast fast test may be performed in disorders where abnormal average RT is observed under a high TP or a fast SOA. For instance, a previous study has described that end-stage kidney disease participants only present abnormal RT under fast SOA54. The ultrafast version of the CVAT may also be helpful in patients with sleep apnea, as abnormal average RT may be independent of TP or SOA16. More importantly, these observations were found even when considering time on task, highlighting the ultrafast version's possible applicability.

Limitations and suggestions for further studies

The present investigation only applied a visual attention task; thus, it would be interesting to explore if similar results are found with an auditory or tactile attention test.

Regarding the sample characteristics, this study mainly included young individuals, so we cannot generalize the findings to children and older adults, who would benefit the most from the ultrafast version of the test. Future studies should also include more females since most participants were male.

Because the ultrafast CVAT only has high TP and fast SOAs, the test cannot be helpful if the disorder of interest affects RT under a low TP or SOA. Other ultrafast versions should be developed.

Test reliability depends on task characteristics. Typically, test reliability is high in self-paced tests, probability because continuous engagement is consistently required in those tests33,55. In contrast, CPTs exhibit significant variability in test reliability, likely due to the greater number of adjustable parameters. Consequently, CPTs may be more vulnerable than self-paced tests concerning reliability. However, they allow much more flexibility in test construction, which is important when the task must be tailored to specific needs commonly find in neuropsychological assessments. We did not examine test reliability in the present study, as this would extend beyond our scope, but we find this is an important point for future follow-up studies.

Conclusions

This study indicates that VRT is not affected by TP, SOA, or time across the 15-min task in a healthy population. We also found a high correlation and a considerable agreement between the 15-min and 90-s CVAT in the VRT variable. Based on these results, the ultrafast version of CVAT might be helpful in screening for disorders where VRT is known to be affected and in conditions where abnormal average RT is observed under a high TP or a fast SOA.