Building Blocks: Studies in Visual Statistical Learning

Statistical learning refers to the ability to extract regularities in our rich, dynamic, and complex sensory environment. Statistical learning has been described as a necessary building block in information processing because it underlies a range of everyday behaviours. The aim of this thesis was to further our understanding of the mechanisms underlying visual statistical learning by investigating the nature of, and factors influencing, visual statistical learning. I conducted a series of empirical studies on visual statistical learning of temporal regularities (regularities that occur over time). Visual statistical learning of temporal regularities was measured using a task comprising two phases: the familiarisation phase and the test phase. In the familiarisation phase, participants completed a selective-attention task, in which regularities were embedded into both the attended and unattended visual streams. Then, in the test phase, participants completed a triplet-discrimination task, which assessed their ability to extract regularities from the attended and unattended visual streams. A four-point confidence-rating scale was also implemented in the triplet-discrimination task to assess participants' awareness or conscious knowledge of the statistical regularities. Using this visual statistical learning task design, I investigated the effects of selective attention and animacy on visual statistical learning (Chapters 4 and 5). In Chapter 4, I investigated whether animacy affects visual statistical learning by manipulating the animacy status of stimuli in the attended and unattended visual streams. I tested four animacy conditions: (i) living things that can self-initiate movement (animals); (ii) living things that cannot self-initiate movement (fruits and vegetables); (iii) non-living things that can generate movement (vehicles); and (iv) non-living things that cannot generate movement (tools and kitchen utensils). In Chapter 5, I extended the earlier study, by manipulating the animacy status of only the stimuli in the unattended visual stream. I tested two animacy conditions: (i) living things that can self-initiate movement (animals) and (ii) non-living things that cannot generate movement (tools and kitchen utensils). I investigated whether there are age differences in the effects of selective attention on visual statistical learning and whether visual statistical learning of unattended information in older adults can be accounted for by stimulus category (Chapter 6). I tested two stimulus categories: (i) highly-familiar line-drawings and (ii) abstract shapes. There were five conclusions derived from this thesis. First, visual statistical learning is modulated by selective attention. Second, visual statistical learning is affected neither by animacy nor by stimulus category. Third, there are age differences in visual statistical learning of attended information but not of unattended information. Fourth, visual statistical learning of attended information is not exclusively accounted for by unconscious knowledge. Finally, visual statistical learning may involve the use of strategies such as naming or labelling of stimuli. This thesis progresses our theoretical understanding of visual statistical learning by presenting insights into how some key factors such as selective attention, animacy status, and ageing can affect visual statistical learning. These insights inform future research by highlighting the importance of taking these factors into consideration when studying visual statistical learning and its importance and prevalence in everyday behaviours