Elsevier

Cortex

Volume 134, January 2021, Pages 333-350
Cortex

Special Issue "Attention & Motor Processes": Research Report
Long-lasting effects of a gaze-contingent intervention on change detection in healthy participants – Implications for neglect rehabilitation

https://doi.org/10.1016/j.cortex.2020.10.013Get rights and content

Abstract

Patients with spatial neglect show an ipsilesional exploration bias. We developed a gaze-contingent intervention that aims at reducing this bias and tested its effects on visual exploration in healthy participants: During a visual search, stimuli in one half of the search display are removed when the gaze moves into this half. This leads to a relative increase in the exploration of the other half of the search display – the one that can be explored without impediments. In the first experiment, we tested whether this effect transferred to visual exploration during a change detection task (under change blindness conditions), which was the case. In a second experiment, we modified the intervention (to an intermittent application) but the original version yielded more promising results. Thus, in the third experiment, the original version was used to test the longevity of its effects and whether its repeated application produced even stronger results. To this aim, we compared two groups: the first group received the intervention once, the second group repeatedly on three consecutive days. The change detection task was administered before the intervention and at four points in time after the last intervention (directly afterwards, + 1 hour, + 1 day, and +4 days). The results showed long-lasting effects of the intervention, most pronounced in the second group. Here the intervention changed the bias in the visual exploration pattern significantly until the last follow-up. We conclude that the intervention shows promise for the successful application in neglect patients.

Introduction

In this study, we explore the effects of a gaze-contingent intervention developed for the treatment of spatial neglect (Ludwig, Schmid, & Schenk, 2020) on a change detection task in healthy participants.

Spatial neglect is a common consequence of stroke, with stronger and longer-lasting symptoms after lesions to the right hemisphere (Gainotti, Messerli, & Tissot, 1972). The symptoms include a spatial exploration bias to the ipsilesional side and an impairment in orienting and reacting to contralesional stimuli (Kerkhoff, 2001).

The exploration bias can be very well studied by looking at eye movements and occurs for example while the patients carry out a visual search (Sprenger, Kömpf, & Heide, 2002) or look at natural images (Ptak, Golay, Muri, & Schnider, 2009). Not only is the contralesional space underexplored, often patients also show an increase in exploration of ipsilesional space, i.e., repeated fixations of ipsilesional targets (Husain et al., 2001; Sprenger et al., 2002). Furthermore, the bias in gaze positions during visual exploration seems to be an excellent measure of neglect severity and thus seems to lie at the core of the syndrome (Kaufmann et al., 2020). The gaze position of neglect patients was tracked while they freely explored pictures of scenes. The averaged horizontal gaze position (indicating the neglect patients’ bias) correlated highly with the Catherine Bergego Scale, a sensitive neglect measure (Kaufmann et al., 2020).

During multiple object tracking, neglect patients with right hemispheric damage have been shown to make more saccades to the ipsilesional (right) side than to the contralesional (left) side (Walle et al., 2019). The fixation distribution showed a right bias in cases of strong neglect, but a left bias in mild neglect, which suggests that biasing fixations to the usually neglected side (left), might compensate for the deficits found in patients with neglect.

This is exactly the aim of the proposed intervention: correcting the exploration bias by biasing fixations away from the overexplored ipsilesional (right) to the underexplored contralesional (left) space. Since neglect patients often have reduced awareness of their own deficits or at least the source of them (Grattan, Skidmore, & Woodbury, 2018), the intervention is designed to work implicitly, i.e., without explicit instructions as to how to explore the visual display. Instead, the redistribution of visual exploration is achieved by giving immediate feedback to the patients in the form of gaze-contingent updating of stimuli. While the patients carry out a visual search task, their gaze coordinates are measured. When their gaze coordinates move to their “preferred” (ipsilesional) side, the stimuli on this side are removed instantly. While exploring the neglected (contralesional) part of the display, all stimuli are present and visible. The rationale of this intervention is that this will cause patients to explore the hitherto neglected side more and the usually preferred side less.

In this study, we will look at the effects of this intervention on visual exploration of healthy participants. This is done to examine whether the intervention – if able to bias exploration behavior in healthy participants – has the potential to work in neglect patients. However, we had to make a small change in the intervention. Healthy participants – while obviously not suffering from biases as strong as those found in neglect patients – demonstrate a small left bias (called ‘pseudoneglect’, Bowers & Heilman, 1980; Nuthmann & Matthias, 2014), which is the opposite of what is typically found in neglect patients (i.e., a right bias). This left bias in healthy participants is most prominent at the start of exploration (Nuthmann & Matthias, 2014; Zelinsky, 1996) and might reflect the right-hemisphere dominance of the spatial attention network (Thiebaut de Schotten et al., 2011). Thus, in this study, the gaze-contingent mechanism operated in the opposite direction: stimuli were removed on the left when gaze coordinates crossed to the left part of the search display. Consequently, exploration of the left side was discouraged and exploration of the right side was unimpaired.

In a previous study, we tested the effects of this intervention on a number of different exploration and attentional tasks, also in healthy participants (Ludwig et al., 2020). The respective task was always measured before and after the intervention and then the results were compared. In two different visual search tasks, we found that the intervention shifted exploration to the side that could be explored freely (the right side) during the intervention. From before to after the intervention, participants increasingly started their search on the right side and overall dwell times (sum of fixation durations in a given area) showed the same behavioral pattern. Thus, we found consistent effects of our intervention on overt spatial orientation tasks, although not on covert attentional orientation (Posner paradigm), and suggested that the intervention might have potential as a new rehabilitation procedure for patients with neglect.

Before moving to the application in patients, we wanted to establish in the current study (1) that the effects of the intervention transfer to a task further removed from the visual search task used in the training procedure, (2) that the version used is the most effective version to achieve the desired behavior modification, and (3) that the benefits of the training can outlast the duration of the training itself.

To address the first point, we chose a task that also requires overt exploration but had further attentional demands compared to visual search tasks. We chose a change detection task in which the change was not immediately detectable due to change blindness (Rensink, 2002).

Change blindness is a phenomenon describing the difficulty of healthy observers to notice changes in visual displays if the change happens during an eye movement, if the presentations of the display with and without the change are interleaved with a blank screen (flicker paradigm), or indeed in any situation that prevents the detection or occurrence of a motion signal that is usually associated with a change (Simons & Rensink, 2005). It has been suggested that change blindness signals a lack in attention and that shifting attention to the place where a change occurs is a prerequisite for change detection (Rensink, O’Regan, & Clark, 1997). This makes this paradigm suited for our purposes since we want to measure whether the intervention unbiases attention or – in case of healthy participants – biases attention to one side of a visual scene. Due to this additional attentional demand, we suppose that a change detection task under change blindness conditions (in our case the flicker paradigm) might be a sensitive measurement for the spatial distribution of attention and hence also for attentional biases.

Furthermore, it differs more from the task in the intervention than the previously used visual search tasks. In change blindness no classical search template is available, i.e., the relevant item to be detected is not defined by a specific feature like color or a conjunction of features. Instead participants must be alert to a change that occurred between the first and subsequent presentations of a visual display. This sets the change detection task apart from classical visual search tasks.

Change detection might thus be a good experimental paradigm for examining the most important function of oculomotor exploration, namely the detection of novel information that ensures an up-to-date representation of our visual environment.

In sum, a change detection task under change blindness conditions provides us with a paradigm that is sufficiently dissimilar to the training procedure to test the transfer of benefits beyond the training task, tests a particularly important function of oculomotor exploration, and presents us with a sensitive measure for attentional re-orientation. The effect of the gaze-contingent intervention on change detection will be tested in experiment 1 (and in the following experiments).

Since the gaze-contingent removal of the stimuli introduces a handicap for the participants and is introduced as a direct consequence of their eye movements, one could classify it as a very mild form of punishment. There is less research on punishment than on reward, especially in the context of research on human attention and perception, but in animals there are indications that intermittent punishment – while producing weaker results than continuous punishment during the learning phase itself – produces a more sustained suppression of the unwanted response after the cessation of the penalizing regime (Azrin & Holz, 1966). We thus devised an adapted version of the intervention in which the gaze-contingent mechanism was not present on every trial of the intervention but only on average on every second trial. Thus we used a variable-ratio schedule (Ferster & Skinner, 1957) in the hope of producing a more pronounced training aftereffect. The effects of this ‘intermittent’ (or ‘adapted’) intervention were then compared with the original ‘continuous’ intervention in experiment 2.

The effects of a single application of an intervention for patients with neglect are often short-lasting. In the case of neck vibration, for example, the beneficial effects disappear quickly (Karnath, 1995). In the case of prism adaptation some are more optimistic (Luauté, Halligan, Rode, Jacquin-Courtois, & Boisson, 2006), but others conclude that to produce lasting effects, prism adaptation also needs to be repeatedly applied (Newport & Schenk, 2012).

In the previous study on our gaze-contingent intervention (Ludwig et al., 2020), we could demonstrate short-term effects, but no conclusion could be drawn concerning the duration of the effects beyond an hour since this was not assessed.

To qualify as a potential rehabilitation procedure, it is critical for our intervention to produce benefits that outlast the duration of our training by at least a few days. To this end, in the current study, we measured the effects of our intervention over several days (up until 4 days afterwards) and examined whether repeating the intervention produced longer lasting effects than a single application (see experiment 3).

Section snippets

Experiment 1 – effects of the gaze-contingent intervention on change detection

The aim of this experiment was to establish whether the effects of the intervention, namely the biasing of visual exploration to the right side (in healthy participants), also transferred to a change detection task under conditions of change blindness, as these tasks have shown to be sensitive measurements of attentional foci (Rensink, 2002).

Experiment 2 – effects of the intermittent intervention

With this experiment and the comparison of the continuous and the intermittent intervention (following a variable-ratio schedule) we want to show whether an intermittent intervention might yield better results, particularly in the extinction phase, i.e., the posttest (in which visual exploration is unimpaired).

Experiment 3 – effects of the repeated intervention and test of their longevity

Experiment 3 was carried out with two aims in mind. First we wanted to find out for how long the effects of the (original, continuous) intervention could be detected. Our second aim was to examine whether a repeated application of the intervention led to longer-lasting effects than a single application.

General discussion

In this study, we wanted to determine whether the proposed gaze-contingent intervention shows promise in several important markers for a good therapeutic outcome: (1) the transfer to a different task, (2) whether we are using the most efficient form of the intervention, (3) whether one can expect it to produce lasting (as opposed to short-lived) effects at least when repeating the intervention. These points will be discussed in turn. Since the intervention caused measurable effects in healthy

CRediT author statement

Karin Ludwig: Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Writing - original draft, Writing - review & editing, Visualization, Project administration.

Thomas Schenk: Conceptualization, Methodology, Resources, Writing - review & editing, Supervision, Project administration, Funding acquisition.

Open practices

Due to organizational reasons, study data and digital study materials will not be available immediately but as soon as possible – and no later than November 2021 – at https://osf.io/xgbt5/.

Declaration of competing interest

The authors declare no financial or personal conflicts of interest.

Acknowledgements

We would like to thank Julia Thaler, Katharina Niamkovich, Vanessa Fetzer, and Johannes Simon Fischer for their invaluable help in data collection. This research was supported by grants from the German Research Foundation (Deutsche Forschungsgemeinschaft: DFG-SCHE 735/3-1 and DFG-SCHE 735/4-1).

References (39)

  • L. Rinaldi et al.

    Reading direction shifts visuospatial attention: An interactive account of attentional biases

    Acta Psychologica

    (2014)
  • D.J. Simons et al.

    Change blindness: Past, present, and future

    Trends in Cognitive Sciences

    (2005)
  • A. Sprenger et al.

    Visual search in patients with left visual hemineglect

    Progress in Brain Research

    (2002)
  • P. Tseng et al.

    Posterior parietal cortex mediates encoding and maintenance processes in change blindness

    Neuropsychologia

    (2010)
  • G.J. Zelinsky

    Using eye saccades to assess the selectivity of search movements

    Vision Research

    (1996)
  • N.H. Azrin et al.

    Punishment

  • D.M. Beck et al.

    Right parietal cortex plays a critical role in change blindness

    Cerebral Cortex

    (2006)
  • A. Bowen et al.

    Cognitive rehabilitation for spatial neglect following stroke

    Cochrane Database of Systematic Reviews

    (2013)
  • D.H. Brainard

    The psychophysics Toolbox

    Spatial Vision

    (1997)
  • View full text