How is retinal information processed?
How does the brain enable us to recognise objects?
How does the brain guide our limbs to manually make use of objects?
These are the questions we seek to answer in this project. In this endeavour, we carry out magnetic resonance imaging (MRI). MRI is an imaging technique that uses a powerful magnetic field and radio-frequencies to obtain detailed cross-sectional images inside the brain. Functional MRI is an application of MRI that measures brain activity by detecting associated changes in blood flow and deoxygenated haemoglobin – as indexed by the so-called BOLD signal. When an area of the brain is in use, more blood flows to that area and more oxygen is released. By presenting a stimulus to the participant, we can determine whether or not the BOLD signal changes in different parts of the brain. If a particular brain area reveals a response in the BOLD signal, we can then infer that this particular brain area plays a role in processing that particular stimulus. We use fMRI to study: 1) how the retina is represented in the brain, and 2) how the brain uses vision for perception and action.
As light enters the eye, it gets refracted through the cornea and the lens such that an inverted image is projected onto the retina. This topographical representation of the inverted image on the retina is maintained throughout the visual system as it is relayed to the thalamus and then to various cortical structures. We map how the retina is represented in the thalamus and the cortex using the “phase-encoding approach” developed by Marty Sereno and colleagues (1995). The approach is based on the principle that when a stimulus is presented in a cyclical manner, the fMRI blood-oxygen-level dependent (BOLD) signal will follow a similar cyclical profile (see part A in the figure below). A movie of a checkerboard stimulus moving across parts of the visual field is played to the participant in a repeating loop (see parts B & C in the figure below). The BOLD signal is analysed using a Fourier analysis to determine where in the visual field the response is the highest. These experiments consistently give rise in each individual person (provided that they are healthy) retinotopic maps similar to those shown in parts D & E in the figure below – each colour superimposed on the brain represents a different part of the visual field.
Goodale and Milner’s Two Stream Hypothesis.
Much of the conceptual framework for our research regarding higher order vision is guided by Melvyn Goodale & David Milner’s two stream hypothesis (1992). According to their model, the ventral stream from the primary visual cortex to the temporal cortex (purple pathway shown below) analyses visual information for the purposes of perception while the dorsal stream from the primary visual cortex to the parietal cortex (green pathway shown below) is used for the online visual control of limb movements. Some of the strongest evidence for this theory arises form patients with brain damage. Damage to the ventral stream results in visual agnosia, which is the inability to recognise objects visually. In contrast, damage to the dorsal stream results in optic ataxia, which is the inability to use vision to guide the limb in space towards an end-point, such as moving the hand towards a three-dimensional object for the purposes of grasping it. Equally important, damage to the ventral stream does not result in optic ataxia and damage to the dorsal stream does not lead to classical forms of visual agnosia. In short, evidence so far from the study of brain-damaged patients supports a double dissociation in favour of the Goodale and Milner two-stream hypothesis.
The neural basis for selecting actions based on concepts.
There are many instances in which people select actions in response to visual stimuli that do not spatially relate to the actions that they specify – such as our limb movements in response to traffic lights while driving a car or the selection of a functional hand posture on a tool. Actions associated with these stimuli are learned and require the brain to first recognise and retrieve the conceptual meaning of the stimulus, which requires the ventral stream, before an action can be executed, which requires the dorsal stream. According to the two stream model described above, these types of goal-directed actions must require an interaction between the ventral and dorsal streams of visual processing – neither of the two streams can accomplish these actions on their own. Our fMRI studies focus more on how the two streams interact than how they are dissociable. Specifically, we are most interested in the neural mechanisms that underlie the execution of a number of skills that depend on learned associations between a visual stimulus and an action (e.g. playing the piano by sight-reading; driving a car while responding to traffic signals; selecting a functional hand posture on a tool; applying the appropriate fingertip forces for lifting an object with a known weight).
Goodale, M. A., & Milner, A. D. (1992). Separate visual pathways for perception and action. Trends in neurosciences, 15(1), 20-25.
Sereno, M. I., Dale, A. M., Reppas, J. B., Kwong, K. K., Belliveau, J. W., Brady, T. J., … & Tootell, R. B. (1995). Borders of multiple visual areas in humans revealed by functional magnetic resonance imaging. Science, 268(5212), 889-893.
Sperandio, I., & Chouinard, P. A. (2015). The Mechanisms of Size Constancy. Multisensory Research, 28(3-4), 253-283.