Research

The detection and recognition of visual objects is a vital skill for our interactions in the world. To achieve it, the brain has to group local image features to global meaningful objects. This perceptual organisation is a challenging operation for the visual system as objects are often camouflaged in cluttered scenes and their image properties (e.g. position, orientation, size) may change as we interact in complex environments. Our work combines behavioural and fMRI measurements to study: a) the integration of global shapes from local features, b) the construction of shape representations selective to combinations of visual features and tolerant to image changes that are critical for object recognition, and c) the role of learning in coherent shape perception. This work will advance our understanding of the link between structure, function and behaviour in the intact brain, provide new insights into the re-organisation and potential recovery of function in the impaired brain, and have potential applications in the design of rehabilitation programmes and artificial vision systems.

Recent behavioural studies have shown that learning can be a key facilitator in perceptual integration for the detection and recognition of objects in cluttered scenes. Further, neurophysiological studies suggest that learning enhances the sensitivity of neural processing. However, little is known about the role of learning in shaping perceptual integration and visual recognition processes across stages of visual analysis in the human brain. In this work, we use human psychophysics and brain imaging to understand the neural plasticity mechanisms that support behavioural improvement in perceptual integration and visual recognition. We ask: How does the human visual brain learn objects in natural cluttered scenes? Does the human visual brain take advantage of natural image regularities (e.g. grouping of elements with similar orientation) that determine the distinctiveness of targets when learning novel objects in cluttered scenes? Does learning facilitate perceptual integration and shape detection in the absence of regularities that usually mediate the grouping of shape contours in natural images? This work will provide a) significant insights into the role of learning in shaping brain functions that mediate key perceptual and cognitive abilities, and b) the foundation for studying the role of learning in visual or cognitive deficits that impair these functions.

The perception of three-dimensional (3D) structure is critical for object recognition, spatial navigation and interactions within our complex visual world. To investigate the neural mechanisms that mediate 3D perception in the human brain, we use multistable stimuli that offer the opportunity to study visual awareness, as they evoke spontaneous alternations between different perceptual interpretations of the same physical stimulus. Recent neuroimaging studies suggest that both occipitotemporal and higher parieto-frontal circuits are involved in ambiguous perception. In this project we examine the neural basis of ambiguous 3D shape perception and the role of previous experience in shaping our perceptual interpretations of the 3D physical world.

Categorisation, our ability to extract abstract information from our experiences in the world and group it into meaningful units (categories), is a cognitive skill fundamental for interpreting our complex environments. How does the human brain learn about the statistical regularities of perceptual experiences that have not been honed by evolution and development? We address this problem by combining behavioural and fMRI-EEG measurements with advanced mathematical approaches (i.e. machine learning). Our goal is to understand the link between the computations that allow humans to make decisions based on adaptive learning and the neural code in the human brain. By parametrically manipulating the physical input (stimulus), we investigate whether we can predict observers’ decisions on individual trails from brain activation patterns.

Interactions in our complex environments entail prompt decisions for successful actions in novel situations. We examine the links between brain structure and function that mediate the ability of older adults to interpret, learn and categorise novel sensory experiences. Despite significant progress in understanding how the human brain ages at the anatomical and cellular levels, much less is known about the relationship between structural and neural changes that underlie the ageing of cognitive abilities and determine an individual’s functional, rather than chronological, age. A core challenge in human cognitive ageing is to understand the mechanisms that lead to rapid cognitive decline in some older adults while others maintain high levels of cognitive performance. This study aims to: a) combine behavioural methods with multimodal imaging techniques (structural MRI, functional MRI, EEG) to relate cognitive performance in ageing to changes in brain morphometry and function, b) develop and validate novel analysis methods for behavioural and imaging data based on advanced mathematical techniques (i.e. machine learning) that provide powerful tools for studying individual variability in cognitive processes across participants. Our findings will advance our understanding of life-long learning and cortical plasticity and have implications for quality of life, early diagnosis and intervention in normal and pathological ageing.

The perception of three-dimensional (3D) structure is critical for object recognition, spatial navigation and interactions with our environment. The horizontal separation of the eyes provides a powerful cue to 3D shape in the form of binocular disparity. To investigate the neural mechanisms that mediate 3D perception in the human brain, we use 3D displays defined by disparity (an example of a 3D stimulus defined by motion is shown for demonstration purposes). We ask which areas in the human brain are involved in the processing of depth information and whether activity in these regions mediates the observers’ perception of 3D shape. Combining disparity-defined stimuli with ambiguous depth figures affords the opportunity to study changes of brain activity in the absence of changes of sensory input allowing us to reveal those regions most closely involved in the perceptual interpretation of 3D shape.

Understanding dynamic events entails the integration of information about form and motion that is critical for fast and successful interactions in complex environments. Despite the ease with which we identify objects in dynamic and complex environments, the computation of meaningful global forms from local image features on the retina is a challenging task for the visual system. We investigate the role of motion in the perceptual grouping and detection of objects in cluttered scenes. We ask whether exposure to motion shapes our ability to detect targets and their neural representation in the human brain. This work will advance our understanding of the brain mechanisms that mediate our ability to recognise objects in natural dynamic scenes.