Community Research and Development Information Service - CORDIS

ERC

BrainReadFBPredCode Report Summary

Project ID: 311751
Funded under: FP7-IDEAS-ERC
Country: United Kingdom

Final Report Summary - BRAINREADFBPREDCODE (Brain reading of contextual feedback and predictions)

The brain constantly anticipates its environment. Computational theoreticians and neuroscientists have cast this predictive function as a potential unified theory of the brain (Bar, 2009; Karl Friston, 2010). The predictive nature of brain processing is thought possible because we have internal models of our environment based on memory, and which influence our moment to moment perception. Internal models are communicated by cortical feedback throughout the cortex, including to sensory areas. Our primary and completely novel objective was to use brain reading techniques to measure predictive cortical feedback processing in the healthy human brain.

We have utilised a variety of parametric paradigms to investigate different characteristics of cortical feedback signals. We have pioneered methods by which to non-invasively measure cortical feedback in human cortex using functional brain imaging and multivoxel pattern classification; we have revealed detailed visual information content contained in cortical feedback signals; we have described where feedback projects to in visual cortex and with what precision; we have learned about the cognitive involvement of visual cortical feedback by use of a dual task design; we have learned about the temporal dynamics of cortical feedback in the context of real-world visual stimulation; and we have investigated into the multisensory context of cortical feedback processing and how it relates to visual feedback.

Based on all of these findings, we have been able to initiate and reinforce international collaborations. The first goal of these collaborations is to connect our findings with a computational approach; cortical feedback processes are currently largely missing from neural network models of the cortex. Our discoveries will inform new algorithmic breakthroughs in artificial intelligence which can capitalise on brain-inspired information processing using cortical feedback signals. The second goal of these collaborations is to investigate the underlying neuronal mechanisms of cortical feedback processing by partnering with groups working with rodent recording and modelling of neuronal processes.

With these collaborations we have escalated and expanded on the findings of my ERC grant. The biggest successes of my ERC grant, aside from 17 publications and over 80 conference proceedings, is the sharing of my paradigms across labs using different techniques and acquiring microscale measures of my cortical feedback signatures in humans, and our contribution to developing deep neural network models of biological vision. These international collaborations helped to secure another large European grant for my lab, from the Human Brain Project. We are part of a consortium taking a multiscale and multispecies approach to develop a theory of the brain which blends bottom-up and top-down approaches to investigate cognitive processes.

Building upon the findings from my ERC grant, we have also started a clinical collaboration with Jiook Cha in New York, for the question about how much emotional and valence content is contained in the cortical feedback signal. Here we are testing if cortical feedback of visual processes is disturbed in individuals with anxiety disorders. Also, in order to understand the developmental trajectory of this feedback mechanism in early visual processing, we are now investigating this in children and adolescents.

These collaborations were the groundwork for structural development within my Host University; we now have a 7T brain imaging centre of excellence. My host institution is equipped with state-of-the-art technology to support these collaborations. In March 2017, we opened our Imaging Centre of Excellence housing a 7Tesla MRI scanner, allowing for highest-resolution imaging of the human brain. The scanner will be furnished with the latest technological advancements in a 32-channel Transmit Receive Head Coil with increased spatial resolution to measure the visual cortex.

Reported by

UNIVERSITY OF GLASGOW
United Kingdom
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top