|
|
|
Conscious & unconscious visionSight Unseen Melvyn Goodale & David Milner, University of Ontario & University of Durham Oxford University Press Keywords: Goodale and Milner, ventral stream, dorsal stream, visuomotor activity INTRODUCTION: Goodale and Milner's work points to a ventral stream in the brain that produces conscious visual perception and a separate dorsal stream supporting non-conscious visuomotor orientations and movements. For consciousness studies, this is important in that it appears to refute the simplistic, but often well-received contention, that consciousness is just what it is to have a brain. Rather this supports the idea that some processes peculiar to parts of the brain or patterns of processing in the brain are the physical instantiation of consciousness. Goodale and Milner refer to their extensive research on a patient, Dee Fletcher, who had suffered brain damage as a result of an accident. Dee has difficulty in separating an object from its background, which is a crucial step in the process of visual perception. She can manipulate images in her mind, but cannot perceive them directly from the external world, thus demonstrating that images generated by thought use a different process from direct perceptions of the external world. In modern neuroscience, perception is not viewed as a purely bottom-up process resulting from analysis of patterns of light, but is seen as also requiring a top-down analysis based on what we already know about the world. On a simple analysis Dee's visual problems were paradoxical. If a pencil was held out in front of her, she couldn't tell what it was, but she could reach out and position her hand correctly to grasp it. This contrast between what she could perceive and what she could do was apparent in a number of other instances. The authors see Dee's state as indicative of the existence of two partly independent visual systems in the brain, one producing conscious perception, and the other producing unconscious control of actions. They point to instances of patients with the opposite of Dee's problems, who are able to perceive objects, their size and location, but are unable to translate this into effective action. These patients may be able to accurately estimate the size of an object, but are unable to scale their grip in taking hold of it, despite having no underlying problem in their movement ability. These opposite problems are taken to suggest partly-independent brain systems, one supporting perception and the other supporting vision based action. It has been found that even in more primitive organisms there can be separate systems for catching prey, and for negotiating obstacles, with distinct input and output paths. This modularity is also found in the visual systems of mammals. The retina projects to a number of different regions in the brain. In humans and other mammals, the two most important pathways are to the lateral geniculate nucleus in the thalamus and the superior collicus in the midbrain. The path to the superior collicus is the more ancient in evolutionary terms, and is already present in more primitive organisms. The pathway to the geniculate nucleus is more prominent in humans. The geniculate nucleus projects in turn to the primary visual cortex or V1. The mechanisms that generate conscious visual representations are recent in evolutionary terms, and are seen as distinct from the visuomotor systems, which are the only system available to more primitive organisms. The perceptual system is not seen as being specifically linked to motor outputs. The perceptual representation may be additionally shaped by emotions and memories as well as the immediate light signals from the environment. By contrast, visuomotor activity may be largely bottom up, drawing on the analysis of light signals, and is not accessible to conscious report. As such this appears little different from the systems used by primitive organisms, while perception is a product of later evolution. Perceptual representations of the external world have meaning, and can be used for planning ahead, but they do not have a direct connection to the motor system. Ungerleider & Mishkin: The authors refer to a 1982 article by Ungerleider and Mishkin arguing for two separate pathways within the cortex. The dorsal visual pathway leads to the posterior parietal region, while the ventral visual pathway leads to the inferior temporal region. The authors relate these two basic streams to the concepts of vision for action and vision for perception respectively. Studies have shown that damage to the dorsal stream results in deficits in actions such as reaching, while the ability to distinguish perceived visual images remains intact. Other studies have shown that damage to the ventral stream creates difficulties with recognising objects, but does not impair vision-based actions such as grasping objects. An interesting study shows that attempts of patients with dorsal damage to point to images actually improved, if they delayed pointing until after the image had been removed. It was surmised that with the image gone, patients started to rely on a memory based on the intact ventral stream. Neurons in the primary visual cortex fire in response to the position and orientation of particular edges, colours or directions of movement. Beyond the primary visual cortex, neurons code for more complicated features, and in the inferior temporal cortex they can code for something as specific as faces or hands. However, while neurons may respond only to quite specific features, they can respond to these in a variety of viewpoints or lighting conditions. By contrast neurons in the dorsal stream usually fire only when the subject responds to the visual signal, such as when they reach out to grasp an object. The ventral stream neurons appear to be moving the signals into perception, whereas the dorsal stream neurons are moving signals into action. The visuomotor areas of the parietal cortex are closely linked to the motor cortex. The authors suggest that the dorsal stream may also be responsible for shifts in attention at least those made by the eyes. The ventral stream has no direct connections to the motor region, but has close connections with regions related to memory and emotions, and is seen as feeding into motor programming, and providing information as to the function of objects. Even within the ventral/perception stream there are separate visual modules. Damage related to any one module can result in localised deficits such as recognition of faces or of landmarks in the spatial environment. A cluster of visual areas in the inferior temporal lobe are responsible for most of these modules. There is also modularity in the dorsal stream, in terms of the type of actions that are guided by particular regions of the parietal, with distinct systems for fast and slow eye movements, hand movements and whole body movements. Blindsight: With the phenomenon of 'blindsight' patients have damage in the primary visual cortex V1. If V1 is not functioning, the relevant visual cells in the inferior temporal cortex remain silent regardless of what is presented to the eye. However, Larry Weiskrantz, an Oxford neuropsychologist, showed that patients with this conditions could nonetheless move their gaze towards objects that they could not consciously see, and later studies showed they could scale their grip, and rotate their wrist to grasp such objects. These blindsight abilities are mainly visuomotor. It is suggested that in these cases, signals from the eyes could go direct to the superior collicus, a midbrain structure that predates the evolution of the cortex, and thence to the dorsal stream. It is suggested that while the ventral stream depends entirely on activity in V1, there may be an alternative route for the dorsal stream. Studies have shown the dorsal stream to be active even when V1 is inactive. The patient, Dee Fletcher, is considered to be similar to blindsight patients, although in her case V1 is still active, which accounts for her still having conscious vision, albeit with impairments. The authors ask why we appear to have evolved two partly independent visual systems. Visual perception provided by the ventral stream is seen as allowing us to plan and envisage the consequences of actions, and to file representations in long-term memory for future use. Motor control on the other hand requires immediate accurate information using a metric that correlates with the external world. In perception the metric is relative rather than absolute. Thus in a picture or film the actual size of the image doesn't matter, and we judge scale by the relative size of objects such as people or buildings. Thus the computation used for the absolute external accuracy of the motor system needs to be different from the relative computation of visual perception. The authors compare this to the present state of robotics, where human operators specify a required action, which the robot cannot determine, but without the operators specifying precise distances or movements, a metric which the robot can provide. Spatiotemporal Dynamics of Synchronous Activity across the Visual Cortex Charles Gray & Baldwin Goodell, Dept. of Cell Biology & Neuroscience, Montana University In:- The Dynamic Brain – Eds. Mingzhou Ding & Dennis Glanzman - Oxford University Press (2011) Neural activity appears noisy and unreliable, but is able to support a complex repertoire of behaviour. Neurons receive inputs from tens of thousands of synapses that are both probabilistic and subject to change. These neurons are organised in networks with a high degree of feedback. The behaviour of individual neurons is characterised by variability of response to the same or continuous signals. Simpler systems outside the brain that are organised on these principles are also characterised by hard to predict outcomes. The authors have used newly developed instrumentation to measure local field fluctuations that show stimulus-dependent synchrony between several areas of the visual cortex. They show that synchronous assembles can form and dissolve rapidly, follow complex trajectories and varying their spatial distribution over a wide range of frequencies. On the basis of a large body of evidence, the authors consider that spatially distributed synchronous patterns are fundamental in cognitive functioning, correlate with sensory input, and are robust and widespread. Cognitive and perceptual processes involve large spatially distributed and synchronous patterns within and between areas of the cortex. These patterns are shifting, transient, occur in multiple frequency bands and are temporally correlated to sensory input and behavioural output. In the period of a few hundred milliseconds, numbers of synchronous networks can form, disappear and reassemble. |
| Email: quantum-mind@blueyonder.co.uk Copyright © Quantum Mind 2012. Web design by 7Soft.co.uk. |