Design of the New Foot Psychomotor Vigilance Test (PVT) for Screening Driving Ability

Abstract

The ability to respond swiftly and accurately to visual stimuli is critical for safe driving. The traditional Psychomotor Vigilance Test (PVT) primarily assesses response time (RT) using finger inputs, but these do not directly evaluate foot responses essential for vehicle control. This study introduces a novel Foot Psychomotor Vigilance Test (Foot PVT) designed to measure the RTs of the foot in response to simulated traffic lights. The Foot PVT integrates a traffic light display interface with a three-pedal system, simulating basic driving conditions. RTs are recorded for three colors (blue, yellow, red) displayed in a randomized order, and the response accuracy is evaluated based on the pedal input. The system also measures correction times for errors, offering insights into a driver’s ability to recover from mistakes. Validation experiments were conducted with eleven healthy younger (25 ± 3 years) and eleven healthy older adult participants (73 ± 4 years). The results showed that the older adult participants (818 ± 84 ms) exhibited significantly longer RTs than the younger participants (725 ± 74 ms, p = 0.016), consistent with age-related cognitive and motor decline. Interestingly, the older participants had fewer false starts, suggesting a compensatory cautious approach to responding. The Foot PVT has the potential to serve as a low-cost, efficient screening tool for evaluating driving fitness, particularly for older adult individuals and novice drivers.

1. Introduction

The Psychomotor Vigilance Test (PVT) was developed by David F. Dinges in 1985 and involves measuring the response time (RT) of the fingers by clicking a mouse at the moment when the elapsed time is suddenly displayed on the screen [1]. It is known that performance on the PVT declines when the subject is sleep-deprived, making it a common evaluation method to objectively observe drowsiness, attention, and alertness [2,3,4,5,6].

Human RT refers to the amount of time it takes for a response to begin after a stimulus—such as a visual, auditory, or tactile stimulus—is presented. It is an important indicator of how efficiently the human sensory system, neural transmission, brain processing, and motor system work together [7,8,9], particularly in activities that require quick reflexes, such as driving. In the context of vehicle operation, RT is essential for avoiding accidents, especially when responding to traffic lights, sudden obstacles, or changes in road conditions.

In recent years, the world’s demographics have been steadily increasing, with a significant increase in the older adult population due to improvements in medical care and longer life expectancies. Globally, life expectancy at birth reached 73.3 years in 2024; the number of people aged 60 and older worldwide is projected to increase from 1.1 billion in 2023 to 1.4 billion by 2030 [10]. This demographic change tends to increase the likelihood of older drivers being involved in traffic accidents due to their reduced RT, so it is recommended that policies and programs regarding older drivers be planned and implemented as part of comprehensive community activities to improve the safety, mobility, and health of older adults [11]. In particular, as of 2024, older adults aged 65 and over make up approximately 29.3% of Japan’s total population, which is the highest proportion of older adults in the world. Furthermore, it is predicted that this percentage will reach about 34.8% by 2040 [12]. The number of older adult drivers is rapidly increasing along with the increase in older adult workers due to the super-aging society, and automobiles are indispensable for older adults [13]. In addition, the increase in traffic accidents due to pedal misapplication is a major concern not only among older adult drivers (aged 65 and over) but also among young drivers (aged under 24) [14]. In response, car manufacturers have integrated advanced driver-assistance systems (ADAS), such as automatic emergency braking and acceleration suppression mechanisms [15,16,17]. However, while these technologies improve vehicle safety, they do not address the fundamental issue of individual reaction ability. Therefore, there is an urgent need for tools that enable drivers to objectively evaluate their own RTs and identify potential deficiencies.

To address this gap, we propose a new application: the Foot Psychomotor Vigilance Test (Foot PVT). This application allows drivers to easily examine their foot RTs by measuring their responses to visual traffic lights. The Foot PVT is a human–machine interface that integrates signal display, vision, foot response, and a foot pedal system. It provides a comprehensive assessment of the psychomotor responses essential for safe driving. Unlike traditional hand-based PVT systems, this application focuses on the foot, which plays a pivotal role in vehicle control. By addressing both the technical and psychological aspects of RT, this work aims to contribute to road safety by improving driver assessment tools. The design and testing of the Foot PVT application, including its interface, operation logic, algorithms, and initial validation results, are described in the next section. The following sections detail the literature review, design, validation, and application of the Foot PVT.

2. Literature Review

RT has been extensively studied in various contexts, including cognitive psychology, neuroscience, and driving safety research. Traditionally, RT has been measured using simple reaction time and choice reaction time tasks, which have been widely used to assess the speed and efficiency of neural processing [18,19,20].

In driving-related studies, RT has often been measured using driving simulators, in which participants respond to simulated traffic events, or through field tests that analyze braking RTs in real-world driving scenarios [21]. Previous studies have demonstrated that slower RTs are associated with poor driving performance and are influenced by factors such as fatigue, distraction, and age-related declines in cognitive and motor function [22,23,24,25].

Previous studies on RT for older adult drivers have shown that older adults tend to have longer RTs than younger drivers due to cognitive and motor decline [26,27,28]. However, older drivers also exhibit compensatory behaviors, such as driving more cautiously, which can mitigate some of the risks associated with delayed RTs [29,30,31]. Additionally, visual perception plays an essential role in RT, particularly among older drivers, who may experience difficulty processing visual information while driving [32].

Several studies have explored different methodologies for assessing driving ability. Some researchers have measured RT in active drivers using devices such as multicolored stimulus presenters, small sensor mats, and driving simulators [33,34]. These studies report individual variations in RT, with younger participants generally responding faster than older participants. Furthermore, factors such as fatigue from long-term driving and mental stress effects have been noted to impact RT measurements [35,36].

Despite the advances in driver assessment tools, there remains a need for an inexpensive, easy-to-use method for evaluating foot RT in a realistic driving context. The Foot PVT seeks to address this need by providing a simple, low-cost yet effective assessment tool that can be used to evaluate both younger and older adult drivers objectively.

3. Design

3.1. Traffic Light Display Interface

This application was developed using Microsoft Visual Studio Community 2015 (Microsoft Corporation, Redmond, Washington, USA) and features a simple user interface with the following components:

(i)

Input Fields: The home screen allowed users to enter a file name to save data and to specify the desired test duration (Figure 1).

(ii)

Traffic Light Display: The colors were arranged from left to right in the same order as Japanese traffic lights: blue, yellow, and red (Figure 2). Before the test began, the background color is black, and three white circles were aligned in the center of the screen. Once the test started, a color was randomly displayed in one of the circles. Although traffic lights typically change in the order of blue, yellow, and red, the order was randomized to create mental strain for the user. The interval before the colors appeared was set between 2 and 10 s.

3.2. Foot Pedal System

Automobile pedals can be classified into two types: a two-pedal system (accelerator and brake) and a three-pedal system (including a clutch). For this study, we adopted a three-pedal foot switch (USB Foot Switch, S23-P, Shenzhen Winsun Technology Co., Ltd., Shenzhen, Guangdong, China) (Figure 3). However, to improve the operability of the Foot PVT, we did not fully replicate the three-pedal configuration of an automobile. Instead, the center pedal was used as a standby pedal when no color was displayed. The right pedal corresponded to blue, while the left pedal corresponded to both red and yellow. The standby pedal did not correspond to any key on the keyboard.

3.3. Strict Application Validation Rules

The instructions/rules for the test were as follows: Before starting, press the center pedal and wait. When blue appears, press the right pedal. When red or yellow appears, press the left pedal. If the color and pedal are correct, the screen will turn black, and you will press the center pedal again and wait. If you press the wrong pedal, the screen will not change until you press the correct one. In the case of a false start, the screen will display “false start” and turn black again, and you will need to press the center pedal and wait for the next display. The time from when the color appears to when you press the pedal is recorded.

4. Build Instructions

4.1. Application Algorithm

Figure 4 shows the application flowchart. First, the CSV file name for saving data and the test time were entered, and Timer1 ran until the end of the test. The time until the color was displayed (interval) was determined by a uniform random number. The range of the uniform random number was set between 2000 and 10,000 ms. Additionally, blue: 0, yellow: 1, and red: 2 were predefined, and the display color was determined by the remainder when the interval value was divided by 3. Once these values were set, Timer2 started and stopped during the interval (no color was displayed). Afterward, the Stopwatch (SW) started the moment one of the blue, yellow, or red circles appeared on the screen and began measuring the RT. The SW stopped when the user pressed the pedal or made a false start. After recording the event results in a CSV file, the event was classified based on the user’s reaction (true, false, or false start). For true cases and false start cases, the interval setting looped back. If the pedal pressed was incorrect, the color remained displayed, and the SW restarted, preventing the user from progressing to the next phase until the correct pedal was pressed. Timer3 was used to show and clear “True”, “False”, and “False Start” on the display. Timer1 monitored the elapsed time and ended the test once the set time was reached.

4.2. Output Data

The application was configured to export data in the CSV format. The CSV file recorded the event number, the color displayed, the evaluation, the interval, the RT, and the elapsed time for each pedal press. From this information, the correction time from the false pedal press to the true pedal press could also be calculated.

To confirm the operation of this algorithm, a PC with the following specifications was used:

• Operating System: Windows 10;
• Processor: Intel(R) Core(TM) i3-4100M CPU @ 2.50 GHz;
• RAM: 12.0 GB;
• System Type: 64-bit operating system;
• Display: 22.4 inch.

Table 1. An example of an output file.

Event No.	Color	Evaluation	Interval [ms]	RT [ms]	Elapsed Time [ms]
1	Yellow	T	9598	691	10,300
2	Red	T	9755	855	20,930
3	False start	F	3880	−915	23,895
4	Yellow	T	6058	820	30,807
5	Blue	T	3210	783	34,837
6	Red	F	7601	1086	43,551
6_Retry	Red	T	0	2125	45,676
7	Yellow	T	3319	1004	50,015

shows an example of the output file. The event number for a false pedal press was displayed as “_Retry”. In the Evaluation column, “T” represented True, and “F” represented False. In this example, the third pedal press was a false start, and the RT was −915 ms. This indicates that the pedal press occurred 915 ms earlier than the interval value. Additionally, when the red light was displayed on the sixth attempt, the false pedal was pressed. The user corrected the mistake by pressing the true pedal. The correction time in this case was recorded as 2125 ms in the RT column of 6_Retry. A one-minute video of the demonstration and the corresponding results are provided in the Supplementary Materials. The results differ from those in Table 1.

4.3. Processing Time

Table 2. The time taken other than interval and RT for each event.

Event No.	Interval + RT (ms)	Time Taken for One Event (ms)	ΔT (ms)
1	10,289	10,300	11
2	10,610	10,630	20
3	2965	2965	0
4	6878	6912	34
5	3993	4030	37
6	8687	8714	27
6_Retry	2125	2125	0
7	4323	4339	16

shows the time taken, excluding the interval and RT, for each event calculated from Table 1. The time taken for one event was determined as the difference in elapsed time for each event. ΔT was the difference between the time taken for one event and (interval + response), indicating the processing time excluding the interval and RT. For the third false start, the result was written to the CSV file the moment the pedal was pressed, so ΔT was 0. For the sixth retry, feedback processing occurred after Timer3 began to run; however, because this time was very short, ΔT was also 0. For all other events, feedback processing was performed to transition to the next event after pressing the true pedal. The processing time for these events ranged from approximately 11 to 37 ms.

4.4. Software Specifications

The software source code can be downloaded from the Supplementary Materials.

(a)

Foot mouse pedal settings: The right pedal corresponded to the right click and the left pedal to the left click of the mouse. The center pedal did not correspond to anything.

(b)

Microsoft Visual Studio Community 2022 was installed.

(c)

The solution file(Foot_PVT.lsn) and the project folder(Foot_PVT) were placed in the same folder and the solution file was opened with Visual Studio Community 2022.

(d)

Form1.vb in the solution screen was the home screen; Form2.vb was the test screen; and Module1.vb was a global variable.

(e)

The start button on the toolbar was pressed to start the build and display the FOOT_PVT home screen. At the same time, the “bin” and “obj” folders were generated in the project folder.

(f)

The output file name and the test time were entered on the home screen and the Enter button was pressed.

(g)

The test screen was displayed, and pressing the Start button at the bottom right of the screen started the test.

(h)

When the test was finished, an output file was generated in the debug folder in the bin folder. The output file name was filled in with the time when the Enter button was pressed on the home screen.

5. Validation

Eleven healthy younger participants (25 ± 3 years old, three females) and eleven healthy older adult participants (73 ± 4 years old, five females) were asked to use the Foot PVT app to verify their RTs. All the eleven younger participants and all the older adult participants held driver’s licenses and were currently driving cars in their daily lives. The younger participants had an average of 3 ± 3 years of driving experience, while the older adult participants had an average of 49 ± 8 years of driving experience. First, the participants performed a 3 min pre-test to familiarize themselves with the operation, followed by a 10 min measurement. The participants operated the pedal with their right foot after removing their shoes. The distance between the PC monitor and the participants was 1.5 m.

Table 3. Younger participants’ Foot PVT App performance.

Participants	Interval [ms]	RT [ms] (Mean ± S.D.)	ΔT [ms]	Correction Time [ms]	Number of False	Number of False Start
Y1	5719 ± 2048	607 ± 104	31 ± 7	1534, 735	2	2
Y2	5917 ± 2264	712 ± 105	27 ± 8	0	0	0
Y3	5846 ± 2262	788 ± 136	29 ± 7	1722	1	2
Y4	6015 ± 2318	779 ± 178	30 ± 7	0	0	1
Y5	5943 ± 2324	817 ± 107	30 ± 6	0	0	0
Y6	5958 ± 2343	758 ± 137	22 ± 10	0	0	4
Y7	6243 ± 2275	778 ± 110	26 ± 8	0	0	0
Y8	5595 ± 2241	592 ± 85	28 ± 12	0	0	9
Y9	6479 ± 2338	720 ± 141	26 ± 10	0	0	0
Y10	5921 ± 2210	645 ± 108	27 ± 9	0	0	0
Y11	6202 ± 2485	774 ± 162	26 ± 10	561	1	0

and

Table 4. Older adult participants’ Foot PVT App performance.

Participants	Interval [ms]	RT [ms] (Mean ± S.D.)	ΔT [ms]	Correction Time [ms]	Number of False	Number of False Start
O1	6318 ± 2272	742 ± 81	30 ± 7	0	0	0
O2	5598 ± 2177	755 ± 105	26 ± 8	0	0	0
O3	6475 ± 2329	880 ± 172	30 ± 7	0	0	0
O4	5767 ± 2283	944 ± 168	27 ± 8	0	0	0
O5	5829 ± 2321	846 ± 125	31 ± 7	0	0	0
O6	5890 ± 2301	863 ± 118	26 ± 8	0	0	0
O7	5977 ± 2171	782 ± 146	29 ± 7	0	0	2
O8	6855 ± 2312	631 ± 170	26 ± 8	566	1	0
O9	5655 ± 2174	788 ± 74	26 ± 9	277	1	2
O10	6007 ± 2389	890 ± 140	28 ± 8	0	0	1
O11	6434 ± 2201	872 ± 122	26 ± 8	0	0	0

show the interval, RT, ΔT, correction time, number of false presses, and number of false starts for younger and older adult participants, respectively. Interval, RT, and ΔT values are shown as mean ± standard deviation. Among the younger participants, participant Y8 had the fastest RT. Participant Y1 had two false pedal presses, participant Y3 had one, and participant Y11 had one. Participant Y1’s first correction time was 1534 ms, and the second was 735 ms. Participant Y3’s correction time was 1722 ms, and participant Y11’s correction time was 561 ms. The number of false starts ranged from 0 to 9. Participant Y8, in particular, had the fastest RT but also the highest number of false starts. The mean interval was stable for all the participants, ranging from approximately 5700 to 6200 ms, with a standard deviation of 2000 to 2500 ms. ΔT remained stable at around 20 to 30 ms, with no significant delay. On the other hand, the older participants had shorter correction times and fewer false starts than the younger participants. The mean interval ranged from approximately 5600 to 6800 ms, with a standard deviation of 2100 to 2400 ms, and ΔT was approximately 26 to 31 ms, which was not a significant delay compared with the younger participants.

Figure 5 shows the mean and standard deviation of the RT for the younger and older adults. The error bars represent the standard deviation. The mean RT for the younger and older adults was 725 ± 74 ms and 818 ± 84 ms, respectively. The standard deviation of RT for the younger and older adults was 125 ± 27 ms and 129 ± 32 ms, respectively. An independent t-test revealed that the mean RT for the older adults was significantly longer than that for the younger adults (p = 0.016). No significant difference was observed in the standard deviation of RT between the younger and older adults.

Table 5. The ratio of the number of times colors were displayed for younger participants.

Participants	Color			Participants by Ratio
	Blue	Yellow	Red
Y1	27	32	36	95 (9.7%)
Y2	33	23	33	89 (9.0%)
Y3	35	30	28	93 (9.5%)
Y4	27	33	27	87 (8.8%)
Y5	30	27	31	88 (9.0%)
Y6	28	28	34	90 (9.2%)
Y7	28	29	31	88 (9.0%)
Y8	32	33	30	95 (9.7%)
Y9	32	26	24	82 (8.4%)
Y10	27	26	37	90 (9.2%)
Y11	25	25	34	84 (8.5%)
Color by ratio	324 (33.0%)	312 (31.8%)	345 (35.2%)	981

and

Table 6. The ratio of the number of times colors were displayed for older adult participants.

Participants	Color			Participants by Ratio
	Blue	Yellow	Red
O1	21	27	36	84 (8.8%)
O2	34	34	26	94 (9.8%)
O3	29	29	23	81 (8.4%)
O4	26	30	35	91 (9.5%)
O5	23	28	39	90 (9.4%)
O6	30	33	26	89 (9.3%)
O7	31	27	25	83 (8.7%)
O8	30	32	28	90 (9.4%)
O9	30	24	24	78 (8.1%)
O10	22	34	34	90 (9.4%)
O11	29	36	24	89 (9.2%)
Color by ratio	305 (31.8%)	334 (34.8%)	320 (33.4%)	959

show the percentage of times that the young and older participants viewed colors, respectively. Approximately 90 colors were displayed over a 10 min period, and a chi-square test was conducted for both the younger and older adult participants to determine whether there was a difference in the percentage of times colors were displayed. No significant difference was found between the two groups (younger adults: p = 0.990, older adults: p = 0.498), indicating that there was no bias in the number of times colors were displayed.

Hiroki Akiyama et al. (2017) used two multicolored stimulus presenters and a small sensor mat to investigate the selective RT of 16 young people to six colors of light [33]. The RTs to the left and right directions were in the range of 615.9 ± 163.9 ms. Droździel et al. (2020) used Lightworks to measure the RT of 15 active drivers of different ages and genders while driving [22]. When the driver kept their foot on the accelerator pedal while the traffic light was on and then pressed the accelerator pedal and the brake pedal, the average RT was recorded as 680 ± 145 ms. Maximilian et al. (2024) measured the RT using a driving simulator in 32 participants (29.3 ± 4.91 years old, 16 females) and found that the mean braking RT was 690 ± 280 ms [34]. The authors’ results showed a certain degree of individual variation in the RT of the subjects. However, the mean RT of the 11 younger participants was 725 ± 74 ms, which is roughly consistent with the previous study, despite the different experimental protocols.

Aging affects cognitive processing speed and motor reaction efficiency, and older adults have longer RTs than younger adults [26,27,28]. The results measured using the app in this study showed that the older adult participants had a significantly longer RT, approximately 1.13 times longer than that of younger participants.

Previous studies have reported that older adults compensate for the decline in cognitive and sensory functions by adopting more cautious driving behaviors, such as slowing down and increasing the distance between vehicles [29,30,31]. Although the developed app only displayed the color of traffic lights, unlike a driving simulator, the older adult participants had fewer false starts than the younger participants, suggesting that they were more cautious in depressing the pedal. Such compensatory behavior may reduce the risks associated with delayed RT in real driving scenarios, and the number of false starts may be one indicator to evaluate compensatory behavior.

Furthermore, it has been suggested that visual perception plays an important role in RT, especially in older adults. Fengxiang et al. (2025) reported that older adult drivers have more difficulty processing visual information while driving, resulting in longer decision-making times [32]. This is thought to be due to the increased processing time of older adult participants for visual stimuli that change over a period of several seconds, which is reflected in the increased RT. The lack of significant differences in the RT standard deviation suggests that, although the older adults respond slower, it has been suggested that the performance of the older participants is not significantly impaired compared with that of younger participants [37]. However, it has been reported that RT measurements using a driving simulator can result in a learning effect [38]. Since this test also involves repeating the same operations many times in a short period of time, it is possible that a learning effect occurs, and the RT become faster.

There are two types of traffic lights for vehicles: vertical and horizontal, with horizontal traffic lights being more common in Japan. Even with horizontal traffic lights, some countries arrange the colors from left to right as blue, yellow, and red, while others arrange them as red, yellow, and blue. Japan follows the former arrangement. In this app, the relationship between the colors of traffic lights and the position of car pedals was crossed, with the right pedal (accelerator) being pressed when the blue light appears on the left side of the traffic light, and the left pedal (brake) being pressed when the red light appears on the right side. When actually driving, the visual field and the traffic light are several tens to a hundred meters apart, so the relationship between the colors and the position of the pedals is not a problem. However, in this application, the distance between the traffic lights and the vision is close, which increases the mental load of associating the color with the pedal position, potentially leading to longer cognitive judgments [39,40]. Therefore, when comparing multiple RTs using this application, it is important to maintain a constant distance between the vision and the display. This mental load can increase the RT in some cases. To minimize such effects and ensure more consistent and comparable results across participants, the users’ display distances should be optimized. Furthermore, the application’s ability to measure and record correction times (e.g., when a participant makes a mistake and corrects it) adds an additional layer of information not typically available from traditional RT tests. This feature is particularly useful for assessing a driver’s ability to recover from mistakes, which is a crucial aspect of safe driving. The application can be refined by considering additional metrics such as fatigue levels, which could further enhance its effectiveness in assessing real-world driving performance.

It is important to realize that this test focuses on a specific aspect of driving ability, namely RT to traffic signals, and does not evaluate other important factors such as decision-making at intersections, judgment under pressure, and situational awareness. Therefore, this tool should be used in conjunction with other driving assessments to more comprehensively evaluate driver abilities. Overall, the Foot PVT application shows promise as a cheap, simple, and effective tool for screening driving ability. However, further studies involving larger and more diverse sample sizes are necessary to validate the reliability of the tool and explore its potential application in various driver screening programs.

6. Conclusions

In this study, we developed a new Foot PVT application as a driving ability evaluation tool. By combining random traffic light displays and pedal operation, this application can efficiently measure the reactivity of the foot, which is difficult to evaluate with conventional PVTs that use fingers. In addition, it is possible to record not only the RT of the pedal press but also the false correction time, which is a feature that can evaluate the ability to recover from errors, an important aspect of driving safety. Since the Foot PVT operates with a simple hardware configuration, it is expected to be used in a wide range of environments for low-cost alternatives. It is considered to be particularly useful in evaluating the ability of older adults, who are easily affected by aging and fatigue, and novice drivers.

As for future prospects, in order to prevent distracted driving, we are considering investigating the relationship between participants’ biological stress, such as drowsiness, physical and mental stress, and activity level and RT, correction time, and number of false starts, and also developing a simple method for estimating fatigue before driving. We hope that the Foot PVT will be widely used as a tool to objectively evaluate driver ability and contribute to preventing traffic accidents and supporting safe driving.