Abstract
Dual-task methodology and Mobile brain/body imaging (MoBI) has been increasingly put in use to understand cognitive motor interference (CMI) while behaving actively in a real-world environment. In the current study, we investigated whether act of walking and displacement through space interfere with visual selective attention using a dual-task paradigm. We assessed young females and males (N = 20; mean age =27.75 years) and compared behavioural and ERP measures associated with attention. Specifically, we measured the P3 effect by employing a visual oddball task during which participants walked on a treadmill, were pushed in a wheelchair, stood and walked in a corridor. By having these conditions, we could separate the independent influence of act of walking and displacement through space. Generally, the classic P3 effect was seen to be present in each of the conditions. A similar P3 effect was observed for act of walking and displacement through space, both in isolation, when interfered with visual selective attention. We expect that access to attentional resources has been integrated indistinguishably for the visual-motor and visual-vestibular mechanisms. An equal reduction in the P3 effect was found for walking, treadmill and wheelchair conditions in comparison to standing condition indicating a competition, for the attentional resources available, among the visual task and motor-vestibular tasks. The reallocation of attention as seen in the neural marker of attention was not evident on a behavioural level. The study maintains that when compared with single-task, in a dual-task, such as walking while performing a visual task, attentional resources are distributed to the secondary task on the cost of primary task. This implies that in a real-world the use of mobile phones in concurrence with real-world locomotion reduces the attention of people, which could place themselves in dangerous situations.
Keywords: Real-world, visual selective attention, cognitive motor interference, dual-task design, P3, EEG, mobile brain/body imaging, walking, standing, displacement through space, vestibular, visual-spatial navigation
Visual Attention and Locomotion: Interaction in the real world.
Mobile brain/body imaging (MoBI), which employs electroencephalography (EEG) integrated in motion, motor, visual capture and other data streams to investigate brain activity while participants actively explore and interact within a natural environment, has made it possible to increase the ecological validity of cognitive science research (Ladouce et al., 2017). Reason for this development is to replicate experimental paradigms, sprouting from traditional setup-based research that contains a highly controllable and standardised environment, in a real-world environment.
One such traditional paradigm that might be useful to apply in a real-world setting is dual-task performance as it examines the attentional demands of motor tasks and the effects of concurrent cognitive tasks on motor performance. Dual-task paradigm have been employed to investigate different aspects of attentional processing by applying various attentional models. For instance, multiple resource pool models or capacity models of attention suggest decreased performance in dual-task situations where two or more tasks compete for attentional resources within multiple resource pools (Wickens, 2002) or the same pool of limited attentional resources (Kahneman, 1973). Primarily, the resource pools are arguably to be utilised by various response channels, signal codes, processing stages, and stimulus modalities (Wickens and McCarley, 2007). By and large, attentional theories position the idea that performance in dual-task conditions deteriorate once the competing tasks exceed the available resource capacity.
Dual-task paradigm has been commonly used to investigate cognitive-motor interference (CMI). CMI demonstrates that a percentage change in motor and cognitive task; together or separately, is the result of concurrent performance of a motor and a cognitive task. Available literature on CMI offers an understanding of utilising the limited resource pools with dual-task paradigm. Janouch et al., (2018) suggests evidence that age-related deficits of multitasking can develop in both concurrent tasks e.g. motor and cognitive tasks implying that the resources available to perform both tasks originate from the same resource pool instead of multiple resource pools and that there is a limited capacity of available resources within the same pool, which is utilized by both motor and cognitive tasks. Such an explanation would contradict the multiple resource model. Therefore, it requires further explanation, especially for ecologically valid scenarios where dual-tasking and multiple tasking is performed frequently.
To navigate in a complex and dynamic environment, a wide range of motor and cognitive abilities are required. On that note, walking, which is considered one of the most common types of human motor activity, is usually coupled with tasks such as spatial navigation, conversations or checking emails on a smartphone (Agostini et al., 2015; Schabrun et al., 2014). For instance, the use of mobile phones in concurrence with real-world locomotion reduces the attention of people to visual inputs (i.e., visual selective attention) and to vestibular and motor inputs in order to constantly calibrate with their environment (Tombu and Jolicoeur, 2003) that could place themselves in dangerous situations.
Attention plays a crucial role in guiding human locomotion and is closely entwined with human motor abilities. This fact could be further illustrated by the means of dual-task paradigms, where both cognitive and motor tasks bring about interference with attentional resources that usually allow Parkinson Disease (PD) patients to compensate for a portion of their locomotor disruptions (O’Shea et al. 2002). It is difficult or even impossible to walk, let alone walk while performing a secondary task without support for patients experiencing cognitive impairments following neurological disorders (Latt et al., 2009; ; Maidan et al., 2016; Iosa et al., 2014; Menz et al., 2003). Generally, the ability to multi-task comes with an associated cost which can disrupt the internal multi-modality (cognitive-motor-vestibular-olfactory) mechanism and interfere with day-to-day task performance in the real world.
There is considerable literature on motor interference with cognitive functions (Al-Yahya et al. 2010) even though vestibular interference with cognitive functions still requires further exploration. In the real world, vestibular inputs are crucial for making sense of one’s environment as most of the time, they remain outside the stream of perceptual consciousness under normal conditions (Faivre et al., 2015). Therefore, vestibular inputs could have an interfering role with the cognition in dual-task paradigms. For instance, vestibular inputs, while acting or behaving in the real world, interfere with attentional processing by competing for the same resources used by visual attention mechanisms for relevant and irrelevant stimuli (Desimone et al., 1995).
It is possible to reflect visual attentional processing by presenting a visual target stimuli which typically elicits a neural response and occurs around 300ms after stimulus onset (P3 Event-Related Potentials). P3, an ERP component, is well-associated with attentional processing in Oddball paradigm and is demonstrated to produce maximal amplitudes to represent the task-evoked P3 component. The P3 effect, i.e. difference of mean P3 amplitude between target and non-target stimuli, demonstrates a clear impact on attentional processing and reflects a distinct attentional neural marker.
In dual-task research, the P3 effect is often used to investigate the extent to which visual selective attention is affected by motor inputs. Dual-task research employs conditions which isolate attentional mechanisms specific to motor inputs. These conditions would extract the independent effect of the motor inputs and of other hidden factors on visual selective attention. For instance, Gramann et al., (2010) detected a P3 feature for the simple act of walking interfering with visual attention. In comparison to steady treadmill walking, the P3 feature for standing still was similar reflecting no greater demands on performance of the visual oddball task for both conditions. The reason for this outcome could be the lack of displacement in space present in both conditions, i.e. standing still and walking on the treadmill.
Displacement through space could impair visual attention because of the interference with distinct modalities using the limited resource pool. In order to navigate through space, vestibular inputs have the important function of maintaining postural equilibrium control in humans, which is essential in walking. Patients with vestibular lesions, damage to inner ear and brain controlling balance and eye movements can have a considerable impact on navigation while restraining the influence of visual and proprioceptive inputs aptly. Specifically, when conflicting inputs disturb the direction of information necessary for navigation (Nashner et al., 1981). Thus, in a real-world environment where extra-spatial conditions are in a constant flux, vestibular inputs could affect cognitive processing of the visual objects.
Displacement through space was assessed to be a main factor reducing the P3 effect, during real-world locomotion and for detecting auditory targets (Ladouce et al., in prep.). This decrease of the P3 effect indicates that fewer attentional resources have been deployed towards the performance of the primary task. This finding suggests that displacement through space leads to the reallocation of attentional resources in order to match the processing demands imposed by the continuous input of sensory information.
Therefore, while (Gramann et al., 2010) did not find any differences between standing still and treadmill walking without displacement through space, the study by Ladouce et al., 2017 suggests that walking with displacement through space indeed results in reallocation of attentional resources. The current study aims to demonstrate that displacement through space in naturalistic walking also plays an independent role in the modulation of visual attention. The idea is that in case of visual modality, the reallocation of attentional resources led by displacement through space will be present because when continuous input of visual information in the real-world increases, task-relevant visual information is more difficult to be discriminated.
1.1 Hypotheses
Due to the prominent role of visual information in guiding human displacement through space (Thomson 1980; Lee 1986), the P3 effect may be observed across conditions reflecting differences in the way visual processing and aspects of displacement through space interact. We speculate that visual and vestibular modalities use the same attentional resource pool and have an adaptive distribution mechanism in order to meet the attentional demands imposed by dynamic environment, while maintaining the performance. Conversely, the adaptive distribution mechanism would help in predicting the attentional resources available once the competing tasks exceed the resource limit.
In the current study, ERP analysis was characterised by measuring the P3 effect at electrode Pz, where the oddball-P3 is eminent (Debener et al., 2002). Specifically, we recorded brain activity of participants performing a visual oddball task i.e. presenting an array of repetitive target stimuli, which are infrequently interrupted by a non-target stimulus.
The visual task was performed under four conditions following a 2×2 factorial design with experimental factors; act of walking and displacement through space. Standing was used as a baseline condition since both experimental factors (act of walking; displacement through space) are absent. In contrast, the natural walking condition involves both the act of walking and the displacement through space. To isolate the contribution of each of the experimental factors i.e. act of walking and displacement through space, two further conditions were tested; During the treadmill condition, participants were effectively walking but remained stationary (act of walking; no displacement through space) while in the wheelchair condition, participants were wheeled down the corridors; hence participants were moving through space but not actively walking themselves (no act of walking; displacement through space). The study investigated the difference in the mean P3 amplitude for target and non-target trials while subjects performed the task in different conditions. The aim is to test how the size of the P3 effect is modulated by the experimental factors of displacement through space and act of walking potentially to an extent.
Based on previous findings, it is expected that the size of the P3 effect will be largest for the standing condition. This is because there is no act of walking and displacement through space present in the standing condition. The size of the P3 effect will be smallest for the walking condition because act of walking and displacement through space are both present in the walking condition. The research will further explore whether displacement through space and act of walking each contribute independently to the modulation of attention to visual targets by looking at each factor in isolation. We predict that; Walking in a corridor, being pushed in a wheelchair and walking on a treadmill each will be associated with a decreased P3 effect on the visual task compared to standing still.
We report behavioural analyses of the participants' accuracy in reporting targets verbal to determine the primary task performance (visual selective attention) in dual-task conditions.
Methods
2.1 Participants
In total, 24 healthy participants were recruited by means of posters and word of mouth at the premises of the University of Stirling. On the whole, exclusion factors concerned participants with motor-visual impairment or a history of psychiatric illness. Four participants’ data were excluded due to considerable noise in the EEG recordings. The remaining 20 participants (mean age 27.75 years ± 7.14 SD; 7 males, 13 females) were familiarised with the task and conditions. University Ethics Panel at the University of Stirling approved the experimental procedures and all participants provided their written informed consent. For their participation in this study, participants were compensated at a rate of £7.5 per hour.
2.2 Stimuli and cognitive task
The cognitive task was a standard visual oddball task, which was presented on a Surface Pro tablet. The task was presented on the tablet running Psychopy (Version 1.85.2; Peirce JW). Target stimuli were represented by a ‘red circle’ while non-target stimuli were presented by a ‘blue square’. The diameter (circle) and side (square) were equal i.e. 300 pixels. The ratio of the ‘Target Vs. Non-target’ stimuli was set at 1:4 to pseudo-randomize the stimuli presentation and additionally, was consistent across all the participants and in each condition. The total number of trials per condition were 300 and the duration of one trial was 1 second.
2.3 Design and procedure
Participants were tested on four 5-minutes experimental conditions and were instructed to count the number of target stimulus (red circle) for the visual task on each condition.
In order to distinguish the contribution of the act of walking and displacement through space on attentional resources deployed towards the processing of task-related stimuli, these factors were manipulated through a factorial within-subject design. Standing was used as a baseline condition since both experimental factors, act of walking and displacement through space, were absent. In contrast, the natural walking condition involved both act of walking and displacement through space act of walking and displacement through space. Two intermediate conditions were used to isolate the respective contribution of the experimental variables investigated. In the treadmill condition, participants were effectively walking but remained stationary (act of walking; no displacement through space). In the wheelchair condition, the participants were wheeled down the corridors. Subsequently, displacement through space was present despite act of walking. Generally, the order of conditions was counterbalanced, and the task was performed only once for each condition. Participants chose a comfortable walking speed at the beginning of the walking and treadmill conditions and maintained their normal speed for their duration.
2.4 Event-related potential recording and analysis
By means of an A 32 channel EEG system (eegoâ„¢sports, AntNeuro, Netherlands), participants’ brain activity was recorded at a rate of 500Hz while sintered Ag/AgCl electrodes were positioned according to the 10 – 20 international reference system. Prior to recording, the electrode impedance was lowered to less than 10 ã€. The P3 wave was recorded as a positive deflection.
MATLAB (Mathworks Inc., Natick, MA) as well as EEGLAB (Delorme & Makeig, 2004) were used to conduct the analysis of recorded data offline. First and foremost, continuous data was visually examined, whereas portions of the EEG displaying levels of noise (e.g. neck and facial muscular activity) and time-intervals between the conditions were manually rejected. While minimal distortion in the recordings was introduced, parameters were fine-tuned to enhance the signal-to-noise ratio of relevant ERP components. Finite Impulse Response (FIR) filters were applied with cut-offs at 1Hz for the high-pass filter and 30Hz for the low-pass filter. The EEG signal epochs were extracted, time-locked from -200 to 800ms relative to visual stimulus onsets and the continuous EEG was split into consecutive epochs of 1 second.
Epochs demonstrating a probability of occurrence > 5 SD from the mean across all epochs were excluded from further analysis. The compromise between discarding further noise and losing higher numbers of trials was upheld by deploying relatively lenient boundaries for rejection thresholds. Taking properties of channel-epoched data kurtosis distribution, absolute power and voltage thresholds into consideration, epochs representing abnormal levels of noise were rejected in the EEGlab toolbox (Delorme and Makeig, 2004). Remaining epochs were averaged to form EEG channel-based ERPs.
Applying to each recording session across all participants, intra-subject differences and potential variability across conditions, which could be rooted in confidence intervals or impedance changes, were computed. Then a two-step process was conducted; the first, an extended Independent Component Analysis (ICA, Bell & Sejnowski, 1996) was performed to obtain Independent Components (ICs) decomposition matrices. Subsequently, ICA weights matrices are back-projected to the original continuous data (post preparation for ICA analysis) to proceed with a less aggressive processing while preserving the ICA decomposition information for rejecting the artefactual components. In the second step, the initial continuous datasets were FIR-filtered to the frequency range under investigation (1 Hz to 30Hz). The current study implemented an averaged mastoids reference for the quantification of P3 ERP.
2.5 Statistical Analysis
Across all participants, the P3 component time-window was defined based on an interval of two standard deviations around the average of single-trials P3 latencies. First of all, the latency at which maximum voltage was recorded was defined for each single-trial, within an extended a-priori window (250-500ms). With these values, the mean P3 ERP amplitude was then computed across channels using a 300-400ms time window. Furthermore, the P3 effect was measured by calculating the difference between the target and non-target trials mean P3 amplitude.
Statistical analyses were conducted using SPSS Statistics for Mac (Version 23.0. Armonk, NY, IBM Corp). Descriptive statistics summarising dataset’s central tendencies (e.g., mean) and their variability (e.g., standard deviation) were performed. For each P3 feature, a 2×2 factorial repeated measures analysis of variance (ANOVA) was performed subject to factors “Act of walking†and “Displacement through spaceâ€. Post hoc analysis was employed for further testing of main effects. Paired-sample t-tests have been applied to the data, effectively providing probability values (p-values) of the null hypothesis (i.e., datasets compared are not different). These p-values were considered as reflecting significant difference between datasets at a threshold of 5% (p < .05), under which the null hypothesis was rejected.
Results
3.1 ERP Results
Supporting our hypothesis, the P3 effect post 300-400 ms stimulus onset was observed at Pz electrode throughout conditions. Figure 1. a) shows the difference between the target and non-target trials mean P3 amplitude, the P3 effect, at Pz while participants walked on the treadmill (Red), stood (Blue), pushed in a wheelchair (Green) and walked (Yellow). Figure 1. b) reflects the largest and smallest P3 effect posteriorly for standing and walking conditions respectively, at Pz across all participants.
The results revealed a significant main effect on the act of walking*displacement through space on P3, F(1, 19) = 8.327, p < .05, ηp2 = .305. In turn, there was a marginal, insignificant variance for the act of walking on P3, F(1, 19) = 4.312, p = .052, ηp2 = .185.
For the displacement through space, there was an insignificant main P3 effect, F(1, 19) = 2.688, p = .118, ηp2 = .124. This suggests that displacement through space alone does not have a significant impact on the P3 effect but together with the act of walking. Furthermore, a series of paired-samples t-test was conducted to compare the P3 effect in the standing, walking, wheelchair and treadmill conditions. Significant difference in the scores for the P3 effect was observed for standing-treadmill, t(19) = 3.0322, p < .05, standing-wheelchair, t(19) = 2.873, p < .05, standing-walking, t(19) = 2.171, p < .05.
The results indicate that the size of the P3 effect was larger for “standing†than for the other three conditions (treadmill, walking and wheelchair) (Figure 2).
3.2 Behavioural Results
The task performance was quantified by computing the difference between the measure (participants’ response) and what was expected (correct response), followed by the absolute difference and finally, the percentage based on this absolute error. The results are presented as follows: Figure 3 shows the mean, standard errors, highest and lowest percent correct response on count of target stimuli during the four conditions: treadmill, standing, wheelchair and walking. Correct response rates were highest for ‘standing’ condition and lowest for ‘walking’ condition.
The data was then statistically analysed using repeated measure ANOVA to compare the effect of factors (act of walking; displacement through space) on percent correct response. The results revealed no main behavioural effect of displacement through space, F(1, 19) = .928, p = .348, ηp2 = .047 and act of walking, F(1, 19) = .038, p = .847, ηp2 = .002.
The behavioural performance results indicated that participants maintained a high level of performance across all conditions suggesting that attention was being deployed primarily on the visual task.
The results indicated that visual selective attention was modulated under the walking, treadmill and wheelchair conditions as compared to the baseline condition i.e. standing. Act of walking and displacement through space modulate visual selective attention under the dual-task conditions (with or without displacement through space and act of walking). Also, the results display a typical P3 topography at Pz (Figure 2) with significant mean amplitude increasing at 300-400ms for the target trials. A substantial difference was found for the P3 effect, between walking, treadmill, wheelchair conditions as compared to standing condition. (Figure 2).
Discussion
In this study, we recorded and analysed brain activity attending to target and non-target visual stimuli (circles and squares, respectively), while participants walked on a treadmill, were pushed in a wheelchair, walked in a corridor and stood. We employed visual oddball ERP paradigm to test whether there was an interference of visual selective attention associated with displacement through space and act of walking. The size of the P3 effect i.e. differences between the target and non-target trials mean P3 amplitude, was measured.
The results demonstrate that we have successfully recorded a P3 response to a visual stimulus. Furthermore, we found that naturalistic walking influences visually evoked response to visual targets. This finding is in line with De Sanctis et al. (2014) which showed walking briskly has a reduced visual P3 amplitude in comparison to while sitting.
Moreover, we observed a decrease in the visual P3 effect whenever the participants not only were walking in a corridor, walked on a treadmill and were pushed in a wheelchair but also when they were standing, which confirms our hypothesis. This suggests that the interference introduced by displacement through space and act of walking in the wheelchair, treadmill and walking conditions, each led to fewer resources available for the primary visual task. This reallocation of attentional resources is reflected in a reduction in the P3 effect. Therefore, active behaviour in the real-world leads to an increase in the vestibular-motor interference with visual selective attention. The increase in vestibular-motor interference appears to be due to both the act of walking and displacement through space. This has not been shown before.
This research is a real-world demonstration of the classic idea that limited resources are available for the primary task (i.e., visual selective attention). To perform the primary task effectively a certain level of attentional resources is required. Therefore, in the absence of competition among the attentional resources, additional resources may have been deployed to the primary visual task. Alternatively, additional resources were reallocated to address the specific demands introduced by the different experimental manipulations (Wickens, 1991).
4.1 Visual P3 Effect: Walking vs Standing
As expected, naturalistic walking (with act of walking and displacement in space) yielded a larger reduction on the P3 effect than standing still (without act of walking and displacement in space). Also, the attentional demands incurred for walking on the treadmill and being pushed in a wheelchair were higher than for standing while performing the concurrent visual oddball task. Therefore, act of walking is independent from displacement through space and it was indeed the displacement through space that accounts for the findings by Gramann et al. (2010). As previously stated, Gramann et al. (2010) found no difference in the P3 feature reflecting visual selective attention in the standing and walking on a treadmill conditions.
The reason for the difference between the current and Gramann’s et al. (2010) findings could be due to divergences in ERP analysis in the previous and current study. In the previous study, the participants were asked to press a button and silently count when they see a target stimulus. In the current study, participants were asked to silently count when they see a target stimulus. Button response to a stimulus can result in the occurrence of specific movement-related potentials. These movement related artefacts could reduce P3 amplitude leading to a decrease sensitivity in attentional processing. Reason for this could be due to an increasing demand of cognitive resources in button-press tasks leaving fewer attentional resources for the visual task. As showed by Salisbury et al. (2001), P3 in an oddball task is smaller for button-press responses than silent-counting.
4.2 Visual P3 Effect: Treadmill vs Wheelchair
Interestingly, the P3 effect for the dual-task in the treadmill condition was similar for the wheelchair condition. This indicates that the visual attention modulated by act of walking was not significantly different from displacement through space. As a result of this, the vestibular and motor inputs consumed the same amount of resources from the limited cognitive resources pool when competing with the primary visual task. It is worth mentioning that behavioural performance was maintained across conditions, suggesting that the minimal requirements in terms of attentional resources allocated to the primary visual task were met. Therefore, the reduction of the P3 effect observed may reflect the reallocation of spared cognitive resources towards the different aspects of locomotion.
Returning to the reasons for similar P3 effect, this could be due to a limited visuo-spatial information. The participants were holding the tablet while they were being pushed on a wheelchair and walked on a treadmill, which may have limited their visual flow. Visual information is essential for real-world locomotion. Therefore, the concentration of visual focus on the tablet may have limited the interference posed on visuo-spatial information.
This research was followed upon from a research in the same lab (Ladouce et al., in prep). In that research, an independent contribution of the displacement through space was found, which is not present in the current study. A difference between the current and previous study is that the modulation of attentional processing during real-world locomotion was investigated in the visual and auditory modality respectively. For this reason, the visual flow during the conditions in the current study might have been limited because of the visual focus on the tablet. In contrast to the previous study, participants detected auditory targets and did not hold any tablet. Therefore, the participants did not have limited visual information and no interference posed on visuo-spatial information.
In the current context of technological development, using a tablet for the presentation of visual information is not the most ecological way to investigate visual selective attention. The presentation of visual information as it occurs in the real-world, holding a tablet restricts the visual field limiting the visual processing of the surrounding space. It is hypothesised that additional visual information related to locomotion would consume more attentional resources from the limited cognitive resource pool potentially resulting in decreased task performance. In the future, to address this issue, the same experimental design could be investigated using an Augmented Reality headset. This apparatus could overcome the above-stated limitations of a tablet by superposing task related visual stimuli onto visual flow during real-world navigation (Krevelen et al., 2010).
4.3 Behavioral Performance
Overall, participants maintained a near-to-perfect level of behavioural performance across all conditions, despite a reduction in the size of the P3. Due to the undemanding nature of the primary visual task, fewer attentional resources were utilised in order to perform the task. To examine the distribution of attentional resources to the primary visual task while maintaining behavioural performance, the level of difficulty involved in the visual task could be increased. By using a three visual stimulus oddball paradigm having target, standard and distractor trials, a higher level of difficulty could be obtained (Hagen et al. 2006). We speculate that in such a scenario, the attentional resources will be utilised in such a way that distractor trials will take away the resources available for the target trials. Therefore, it would be difficult for the participants to attend and count the target trials. This would decrease the behavioural performance of the participants across all conditions.
4.4 Future Implications and Studies
The mechanisms of attentional resource, utilised by the motor and vestibular inputs, could be further investigated by manipulating the wheelchair and treadmill conditions with respect to cognitive load (visual task) or vestibular and motor load (conditions) (Heuer, 1991; Patel et al., 2014). The experimental design in the future studies could vary from condition to condition, according to vestibular-motor and visual task. For the vestibular and motor performance, this could be achieved by employing a 2×2 factorial design where the conditions (treadmill, wheelchair) are measured by increasing the speed i.e. slow and fast while the participants are performing the visual oddball task. We speculate that the increased speed could reveal different pattern for utilising the attentional resources for both vestibular and motor conditions. Previous studies have shown that walking speed and postural control utilise a considerable amount of attentional resources (Woollacott & Shumway-Cook, 2002; Camicioli et al., 1997; Brauer et al.,2002).
To investigate the cognitive load task, the utilisation of attentional resources can be measured by increasing the difficulty of the task i.e. low difficulty and high difficulty for both wheelchair and treadmill conditions. This would reveal whether cognitive load counts for the difference for motor and vestibular inputs in modulating visual selective attention. This increase in the task difficulty would lead to an increase in the reallocation of attentional resources for the visual task. Hence, spared resources will be utilised by vestibular or motor tasks, which would demonstrate whether motor or vestibular inputs are competing more for the available resources.
Further studies could also investigate the P3 ERP components in dual-task paradigms when performing a motor-vestibular task but with old and impaired participants. The aim would be to examine whether the interference can be problematic if an individual’s resources are depleted on the ground of old age or impairment (Haggard et al., 2000; D'Amato et al., 2012). By examining the old and impaired participants, it may be observed that vestibular-motor interference with visual attention leads to reduced P3 effect as compared with young healthy participants. We may observe this reduction due to the limited resource pool. Owing to old age and impaired patients, the sensory and motor processes increasingly require control over the attentional resources due to the deterioration of overall attentional processing (Lindenberger et al. 2000).
The current research did not explore the P3 topography for interference of displacement through space and act of walking with visual attentional resources. It would be interesting to analyse the current study in the future to investigate frontal (Fz) and central-central (Cz) electrodes and hemispheric differences. For instance, Goodin et al., 1985 demonstrated that there are hemispherical differences in P3. Specifically, P3 wave was more prevalent in the left hemisphere when participants discriminated shapes. In dual-task conditions, the P3 effect for visual target stimulus could be larger than for visual non-target stimulus over the frontal region due to the distribution of attentional resources over both tasks and primarily attending to the visual task.
4.5 Conclusion
This study investigated the interference of visual selective attention with the act of walking and displacement through space. The present findings highlight the modulation of neural markers reflecting attentional processing during real-world locomotion for the visual modality. The findings suggest that both the act of walking and displacement through space are main factors driving the reallocation of visual attentional resources. We found that there was a decrease in P3 effect for the treadmill, wheelchair and walking conditions in comparison to the standing condition. This study confirms that in dual-task conditions reallocation of resources appear once the competing tasks exceed the available resource capacity from the same pool of resources. In the current study, it was demonstrated that naturalistic walking modulates visual selective attention. Therefore, the competition for processing attentional resources in the dual-task conditions was present within the general capacity limitations. Potentially, motor and vestibular inputs might have adaptive distributive mechanisms for attentional resources, which predicts the availability of resources once the competing visual task exceeds the resource limit.