Abstract
Perceived speed slows during self-controlled head movement: Testing a Bayesian analysis of motion in both vision and hearing
School of Psychology, Cardiff University
Email:
A hallmark of standard Bayesian models is that prior information increasingly dominates perceptual biases as sensory signals become less precise. Thus, in motion perception, a slow-motion prior explains why targets moving in the dark appear slower when pursued: the extra-retinal signal is known to be less precise than the visual signal present during fixation. We investigated whether the same perceptual slowing occurs during self-controlled head movement, using both visual and auditory stimuli. Alongside, we tested the Bayesian explanation by developing a method for measuring the precision of the underlying head-movement signal, made difficult because head movements are under participant control. Participants sat at the centre of an LED/speaker ring. In a ‘head-movement’ condition, they judged whether a head-fixed target (I.e., moving ‘with the nose’) moved more during self-controlled head rotation (interval_1) than when the head was stationary (interval_2). Motion in interval_2 was based on a scaled version of head-movement recordings in interval_1. In a ‘head-always-stationary’ condition, similar scaled recordings were used, again in a two-interval task, to determine the precision of visual and auditory signals. We then estimated the precision of the head-movement signal by combining the two conditions using signal-detection theory. For both visual and auditory targets, we found perceived speed slowed by around 30% with the head moving. We also found that the precision of the head-movement signal is similar to the precision of visual or auditory signals. These findings are difficult to explain together within a standard Bayesian framework with unbiased sensory inputs.
The expectedness of semantic scene-content as a factor influencing human gaze behaviour
1School of Biological and Behavioural Sciences, Queen Mary University of London, London, UK
2Psychology Department, Northeastern University, Boston 02115, USA
3LIRIS, CNRS, University of Lyon, France
Email:
Visual information in the real world is typically continuous over time. In a pre-registered study, we investigated how this visual continuity influences eye movements, by comparing viewing behaviour when this continuity is preserved and when it is disrupted. Participants freely viewed 80 sequences of image frames from movies. Each sequence contained several ‘context’ frames followed by a ‘critical’ frame. Context frames were related and showed a consistent unfolding of events over time. The critical frame was either a natural continuation of these events (expected condition) or was completely unrelated to them (unexpected condition). We compared five characteristics of eye movements between the expected and unexpected conditions: number of fixations, saccade amplitude, mean fixation duration, probability of blinking and first-saccade latency. Results show that number of fixations and saccade amplitude increased in the unexpected condition relative to the expected condition whereas mean fixation duration and probability of blinking decreased. There was no significant effect for first-saccade latency. These findings suggest that how people look at identical scenes depends on the preceding temporal context, with participants displaying more exploratory visual behaviour and increased alertness when the current visual input is surprising (e.g. unexpected) in the context of preceding visual images.
Facilitation of nonsingleton distractors explains away singleton distractor “suppression” in the capture-probe task
1Institut für Psychologie, Universität Leipzig
2School of Philosophy, Psychology and Language Sciences, University of Edinburgh
Email:
In the capture-probe task, participants report probe letters that occasionally appear during a visual search task, providing a separate attentional measure of each search array item. In multiple studies, Gaspelin and colleagues (e.g., Gaspelin et al., 2015 Psych Sci 26(11)) used the capture-probe task and found reduced probe recall associated with singleton distractors, relative to nonsingleton distractors. This “below baseline suppression” was used as behavioral evidence supporting the signal suppression hypothesis: that capture by a singleton distractor can be preempted through top-down inhibitory mechanisms when there is foreknowledge of its features (i.e. color or shape). Here we tested an alternative account: given that nonsingleton distractors always shared the same color as the search target, while singleton distractor shapes were of a different color, the original capture-probe effect could have been driven by a global feature enhancement of the target color rather than suppression of the singleton. In three experiments, we presented a capture-probe task and replicated the effect seen in the original studies. But when the colors of nonsingleton distractors were changed such that they no longer matched the target color, probe recall difference between nonsingleton and singleton distractors was abolished. We then manipulated the color similarity between targets and nonsingleton items, and found that probe recall for the latter was enhanced when these were more similar to the target. The findings strongly suggest that global target-color enhancement for nonsingleton items can explain distractor performance in the capture-probe task, without the need to appeal to a preattentive suppression mechanism.
Visual conduction delay: the Higg’s boson of decision
1CUBRIC – School of Psychology, Cardiff University
2School of Psychology, Aston University
Email:
Reaction time is a cornerstone of experimental psychology, and its measurement and interpretation are fundamental to vision science. In fast visually-guided decisions, reaction time presumably reflects the sum of three components: sensory conduction delay, decision time and motor execution time. The idea that information is not immediately available for decision but must travel through the brain is, like the Higg’s boson in particle physics, a most useful hypothesis, allowing models to successfully capture empirical data. However, measuring sensory conduction delay behaviourally appeared beyond reach. Recently, we described a method to do this, that does not rely on modelling assumptions, but only on behavioural data and logic (Bompas, Campbell & Sumner, 2020 Psychological Review 127(4): 542-561). This method applies to visually-driven fast action decisions, pervasive throughout the cognitive control literature, on humans and animal models. We gather 12 datasets from multiple labs to outline the defining properties of sensory conduction delay in saccadic and manual responses. We provide a first systematic description of the effect (or lack thereof) of well-documented factors, such as visual properties, speed-accuracy tradeoff or action modality. These effects are contrasted with the conclusions reached from simple decision models relying on default assumptions for non-decision time, i.e. chosen out of mathematical or computational commodity rather than biological plausibility. This work highlights the need for a more rigorous characterisation of non-decision time, alongside the conceptualisation of decision as a dynamic integration of bottom-up and top-down signals.
Investigating the effects of spatial frequency in visual search using equivalent noise
QinetiQ Ltd, Farnborough, UK
Email:
Detection of a target is thought to be limited by internal noise, which disrupts the target signal when the signal is weak (Faisal, Selen, & Wolpert, 2008 Nature Reviews in Neuroscience 9 292-303). One way of testing target sensitivity is to apply external noise, known as the equivalent noise method (Pelli & Farell, 1999 JOSA A 16 647-653). Previous research has found that applying external noise can successfully determine the contrast sensitivity function (Baldwin, Baker, & Hess, 2016 PLoS One 11). When using natural images in a visual search task, the target stimulus is limited in signal intensity range. Therefore, the present study aims to develop a methodology that overcomes this limitation when investigating the effects of spatial frequency of a target. We examine whether applying external noise in a visual search paradigm will lead to reduced sensitivity to higher spatial frequency stimuli, reflected by contrast in the noise mask. To investigate this, observers were presented with a normally distributed noise mask added to a target Gabor stimulus in a visual search task where spatial frequency was varied. From our pilot data, we can see that spatial frequency impacts the sensitivity of the noise mask in a visual search paradigm. We aim to replicate the equivalent noise method in conjunction with a visual search paradigm to validate its use in visual search tasks using natural images.
Taking face space to the extreme: assessing the effect of hyper-caricaturing faces on the fMRI response in the FFA
1 School of Psychology, University of Nottingham
2School of Computer Science, University of Nottingham
Email:
An aspect of ‘Face space’ (Valentine, 1991The Quarterly Journal of Experimental Psychology Section A 43 161-204) that has not been widely researched is the effect of caricaturing on neural responses. Theoretically, a divergence from average should increase neural responses. Some prior research suggests this is not the case, but few studies have taken caricaturing to the extreme. The current experiment sought to test this using computer-generated faces of varying caricature level made from a PCA-based face space. In an event-related fMRI experiment participants were presented with faces of varying naturalness, ranging from fairly average to very caricatured in appearance. Initial findings in the right FFA showed a significantly greater response to the most, compared to the least, caricatured. Interestingly, we found no significant effect of, nor interaction with, upside-down inversion of the stimuli. We performed a within-subjects one-way ANOVA in the right FFA assessing all 5 caricature levels of upright faces. We found a generally monotonic increase in response amplitude with caricature level. There was however a notable peak in the fMRI response to the second level of caricature, which corresponded to the point in a preceding behavioural experiment at which the participants rated the faces as being on the edge of natural, physical plausibility. Overall, there is evidence to suggest increasing neural responses with caricaturing in the right FFA. While this might be explained by low-level visual differences, there may also be an additional effect reflecting naturalness or processing at a categorical boundary.
Multisensory integration of audio-visual motion cues during active self-movement
School of Psychology, Cardiff University
Email:
Sensory measurements in vision and hearing begin in different coordinate frames, which complicates the integration of spatial cues when head and/or eyes move. We propose that audio-visual integration in active perceivers is performed in body-centred coordinates: specifically, audio and visual cues are separately combined with self-movement signals, before being integrated as body-centred cues to audio-visual motion. We tested whether body-centred audio and visual cues are optimally integrated during active self-movement. In all conditions, participants rotated their head back-and-forth while sitting in the centre of a speaker/LED ring. They fixated a head-centred fixation-point to minimise eye rotation, and judged the left/right direction of a moving audio, visual, or audio-visual target. The target’s speed was scaled by a proportion of the ongoing measured head movement. In some conditions, audio and/or visual positional jitter was added to the target’s motion to make cue(s) less reliable. Estimates of precision and bias were obtained via psychometric functions, with parameters from audio and visual conditions used to predict optimal audio-visual estimates. Reducing the reliability of one or both cues caused dynamic re-weighting of sensory information, with a higher weighting given to the more reliable modality. The results suggest optimal integration of vision and hearing while moving, but based on audio and visual cues that are body-centred and partially correlated by virtue of the common self-movement signal they share.
The motion silencing effect in static and dynamic orientation change detection
1School of Psychological Sciences, University of Bristol
2Perception in Action Research Centre, Department of Cognitive Science, Macquarie University
Email:
Crowe et al, 2021 Cognitive Research: Principles and Implications 6
Research Culture and the AVA: how should we engage?
School of Psychology and Neuroscience, University of St. Andrews
Email:
Funders, employers and the academic sector are becoming increasingly engaged in improving our Research Culture (RC). The RC agenda includes a broad swathe of activities including: career pathways, values, attitudes, behaviours and influences the way in which we conduct research. We recently conducted a RC survey at the University of St Andrews, which compliments similar surveys at universities such as Glasgow, Edinburgh, and an international one from The Wellcome Trust. In this talk I would like to summarise some key themes that are emerging from RC surveys across the sector, including bullying & harassment, structures of promotion and recognition, career instability and the influence of external factors such as funding models and the REF. I will highlight some of the interventions that are beginning to be applied and will pose some questions around Research Culture that could be addressed by the AVA community. My talk is an invitation to start a conversation about how AVA could support RC-improvement activity.
Contrast sensitivity changes at different background brightness levels in patients before and after cataract removal surgery
1Optometry department, University of Latvia
2Institute of State Physics
Email:
Contrast vision tests can be a useful method to determine if the surgery is needed and the tests help to understand patient’s complaints about daily life tasks like driving (e.g. Karatepe et al., 2017, Turk J Ophthalmology, 47(2), 80 – 84). Our aim was to estimate the contrast vision sensitivity at different background levels and compare light scattering in patients before and after cataract surgery. Methods: In our research participated 72 patients with cataract and 56 control group patients. The contrast sensitivity was measured with alternative forced choice test design (AFC) before and two weeks after Femto laser cataract surgery. Contrast vision measurements were performed under mesopic conditions at different background brightness levels (60 cd/m2, 85 cd/m2, 100 cd/m2) and spatial frequencies (4 cpd, 6 cpd, 12 cpd, 18 cpd) (e.g. Smedowski et al.2015, Med Princ Pract 2015, 24, 501-508). Results: Cataract-induced light scattering significantly decreased contrast sensitivity at all spatial frequencies. There were no statistically significant differences between the Weber constants at the different background lighting levels, between before the cataract surgery group and the control group. At the lighting level of 60 cd/m2, cataract surgery provided significant improvement at the average spatial frequencies.
Dazzle camouflage one hundred years on: Benefits and problems revealed
1College of Health and Life Sciences, Aston University, Birmingham, UK
2Division of Psychology and Forensic Sciences, Abertay University, Dundee, UK
Email:
During WW1, ships were painted in high contrast ‘dazzle’ patterns of irregular stripes and shapes, presumably to confuse submariners about target heading. One plausible intent was to use pre-Gibsonian texture gradients to induce illusory spatial slant. Some designs (e.g. Wilkinson’s scheme for RMS Mauretania (launched, 1906)) suggest this, whereas others (e.g. Kerr’s designs), do not. Here we report a direction matching experiment to investigate observers’ perceptions of heading for a computer model of Mauretania dressed in various camouflages (Uniform grey, Wilkinson, Kerr, synthetic gradients) for simulated directions around the clock. Our results and analysis confirm that texture gradients ‘twist’ the perceived direction of the ship, the effects being greatest for our own synthetic design. However, computer simulations of torpedo strikes show that the twist deception might have saved fewer lives than often supposed. This owes to a second, hysteresis, effect (easily confirmed by casual observation) where the perceived directions of ships ‘stick’ to the horizontal (left/right, or 270/90 deg) for a range of about +/-25 deg, thereafter paralleling, not matching, veridicality (i.e. a constant pull of 25 deg towards horizontal for much of the remaining range). The asymmetric twist and rotationally symmetric hysteresis combine constructively or destructively, depending on the relative directions of the gradient and of travel. Thus, the deceptive hysteresis is undermined by the twist as often as being enhanced by it. Only if submariners were immune to the hysteresis effect would a dazzle-imposed twist illusion be of substantial value. We discuss the veracity of this striking conclusion.
The effect of temporal frequency on habituation in migraine
1NTU Psychology, Nottingham Trent University
2School of Psychology, University of Lincoln
Email:
Individuals with migraine tend to experience discomfort when viewing flickering visual stimuli. It has been suggested that one of the characteristics of migraine is a lack of habituation to repetitive visual stimuli (e.g. Coppola et al., 2009 Neurobiology of Learning and Memory 92 249-259), although findings can be mixed. Previous work has typically used similar visual stimuli (chequerboard) and only one temporal frequency. This study systematically varied the spatial and temporal characteristics of the visual stimulus, using steady-state visual evoked potentials to assess the differences in amplitude between migraine and control group over consecutive blocks of stimulation. At 3Hz stimulation, there was evidence of increased habituation in the migraine group. However, at 9Hz stimulation, there was evidence of increased potentiation in the migraine group. Visual discomfort and SSVEP response amplitude were affected by spatial and temporal content of the stimuli. This difference in SSVEP response behaviour, dependent on temporal frequency, is important to consider when researching the effects of repetitive visual stimulation in migraine.
The nature of visual discomfort experienced by individuals with and without migraine
1School of Psychology, Cardiff University, Cardiff, UK
2Cardiff University Brain Research Imaging Centre (CUBRIC), School of Psychology, Cardiff University, Cardiff, UK
Email:
Migraine is characterised by intense headaches that are often accompanied by subjective sensory sensitivity (e.g., to sound, light, movement). We were interested in whether such sensitivities are also experienced between attacks and found this to be the case across modalities. Mediation analyses determined that anxiety symptoms acted as a partial mediator to this relationship, although sensitivities also existed independently of anxiety. In our subsequent research, we have taken a symptom approach to consider the role of visual sensitivity and/or discomfort specifically, in individuals both with and without migraine. We are interested in whether visual discomfort differs in kind or in severity across individuals, and how these patterns may associate with different stimuli and behaviours experienced in everyday life. We developed an online questionnaire to investigate these questions. Preliminary cluster analyses suggested that individuals susceptible to high visual discomfort can be differentiated based on their perceived functional impact and scores on measures of visually induced dizziness. Although experience of migraine is relevant to these distinctions, our sample suggested a subgroup of individuals who experience very high levels of visual discomfort and visually induced dizziness (notably, at clinical levels), and report very few neurological conditions. In ongoing analyses, we are investigating the differences in nature of response to visual stimuli known to be problematic for those with high visual sensitivity across these clusters.
Limited evidence for the perceptual reality of synaesthetic colours
1School of Psychology, University of Southampton
2Faculty of Psychology, UniDistance Suisse
Email:
No study has unambiguously shown that synaesthetic colours are perceptually real rather than just conceptually represented in memory. Here we tested for the perceptual reality of synesthetic colours using an achromatic adjustment task. This task is particularly sensitive for measuring perceived colour, and it has been successfully used in the past to show the influence of memory on colour perception. In this study, 20 grapheme-colour synaesthetes were asked to adjust the colour of their individualised synaesthetic inducers (letters and numbers) to grey. Results showed that synaesthetes saw synaesthetic inducers in their concurrent colour even when they were shown without colour (i.e., greyscale). This was not the case for a control group of 30 non-synaesthetes. We also included a control stimulus that reproduced the known effects of memory on perceived colour in both groups of participants. Additionally, we conducted a synaesthetic variant of the Stroop task to replicate previous findings about semantic effects of synaesthesia. Results confirmed that semantic colour associations are automatically triggered by synaesthetic inducers, but not by control stimuli and not in non-synaesthetes. Taken together, the present results suggest that grapheme-colour synaesthesia involves real perception of concurrent colours. However, the colour synaesthetes see in achromatic inducers are very desaturated, even more than those of memory colour effects. This contradicts rather than supports the idea that synaesthetic colour experiences involve vivid perceptual experiences of colour.