Johnston, R.
(2017)
Motion and form vision in typical and atypical adult readers.
PhD thesis, University of Nottingham.
Abstract
Poor reading in adults is often associated with developmental dyslexia, especially for adults that have average, or above average, intelligence and no history of ocular ill health, social deprivation, or other learning difficulties. Developmental dyslexia is primarily associated with poor phonological decoding skills, but evidence suggests that there is also a visual perception deficit. In particular, it has been reported that individuals with dyslexia have difficulty integrating local motion, but not local form, cues into a global percept. However, current theories proposed to explain this visual impairment, cannot fully accommodate the perceptual abilities of readers with dyslexia. The aim of this thesis is to investigate the origin of visual impairment in poor readers, including those with dyslexia.
It is unknown if the global motion deficit in readers with dyslexia reflects a difficulty processing motion, temporal information, or integrating local visual cues across multiple (>2) dimensions. To address this issue four diagnostic, global motion and form tasks were administered to a large sample of adult readers. Results showed that generally poor readers, and individuals with dyslexia, had significantly higher coherence thresholds than relatively good readers on a random-dot global motion task, a spatially 1-D global motion task, and a temporally-defined global form task. In contrast, no difference was found on a static global form task. Thus, the visual impairment in developmental dyslexia represents a difficulty integrating cues that change over time, rather than a difficulty processing motion per se.
The visual system has to satisfy competing constraints of integrating features that belong to a common object, whilst segmenting those arising from different objects. To explore if the visual deficit in developmental dyslexia extends to spatial segmentation, a motion-based spatial segmentation task and an analogous form task was administered to thirty-eight adult readers. Thresholds were measured for each of a range of segment sizes to characterise task performance. Participants were unable to perform either task when segments were smaller than a spatial resolution (acuity) limit that was independent of reading ability. Thresholds decreased as segment size increased, but for the motion task the rate of improvement was shallower for the poorest readers and the segment size at which performance became asymptotic was larger. Thus, the visual deficit in developmental dyslexia also extends to spatial segmentation when tasks contain time-varying (e.g. motion) information.
To explore why motion-based segmentation is impaired in generally poor readers a biologically-inspired computational model was developed. Simulations showed that even though the size of the motion-boundaries in the image was fixed, and predictable, within a single block of trials relatively poor readers could not utilise this knowledge to select a detection mechanism of the appropriate size (scale) for optimal performance. A prediction of this novel “scale selection hypothesis” is that relatively good readers should exhibit a qualitatively similar pattern of performance to generally poor readers, when scale uncertainty is introduced into the task. This prediction was tested empirically and the results were consistent with the scale selection hypothesis.
Eye movements and fixations are made during reading so the positions of words on the retinae change over time. To encode text effectively, this temporal stream of visual information must be segmented over time, as well as across space, but it is unknown if temporal segmentation is also impaired in developmental dyslexia. To address this issue temporal versions of the spatial segmentation tasks were administered to thirty-eight adult readers. Coherence thresholds were measured at each of a range of oscillation rates (segment durations). Results showed that temporal acuity for both the motion and form tasks were significantly and negatively correlated with reading skill. Thus, generally poor readers have difficulty segmenting temporally changing motion and form cues, particular at the briefest segment durations.
An alternative theory of the origin of visual impairment in developmental dyslexia is the anchoring deficit hypothesis. It proposes that individuals with dyslexia cannot utilise stimulus repetitions to form a perceptual anchor, to reduce memory load when comparing sequential stimuli. This theory has not yet been tested using global motion and form tasks. To address this issue a global motion detection task and an analogous form task were administered to forty-four adult readers. By varying the degree of uncertainty associated with the axis of global motion or orientation, the ability to form a perceptual anchor was manipulated across different blocks of trials. Results showed that reading ability was not significantly associated with performance on the global motion task or the global form task, challenging the anchoring deficit hypothesis.
In summary, the results of this thesis are difficulty to reconcile with previous theories of vision in developmental dyslexia. They show that the visual deficit in generally poor readers and individuals with dyslexia impacts on the ability to both integrate and segment local cues in the image and has at least two distinct components, one spatial and one temporal. That is, 1) a difficulty in selecting the spatial scale optimal for task performance; and 2) a difficulty judging the global direction or orientation of stimuli changing rapidly over time.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[img]](/style/images/fileicons/application_pdf.png)