This year at Capo Caccia we will be talking about what forms of computation exist in cortex and how they are implemented, exploring whether competition in cortex can explain high-level representations of visual objects, and thinking about ways to maintain state with cortical oscillations.

Competition in columnar cortex

Over the last two years in capo caccia, we have proposed and developed a new network architecture for winner take all competition in cortex. We wanted to explore network architectures that are faithful to cortical anatomy, and understand how competition can emerge in cortex. We've so far used this architecture to explain decorrelation and sparse activity in columnar cortex, and to implement gestalt visual properties. This new architecture is exciting because it seeks to explain several competition within a cortical column, while exploring the effects of connections between columns. (See the workgroup pages for 2011 and 2012 for details.)

There are many avenues that could be explored this year, depending on the interests and expertise of the participants. We could aim for an analytical characterisation of the computational properties of a simple network. We could explore how the basic local architecture can arise through learning, or explore how learning can adapt long-range functional connections. Alternatively, we could examine how top-down inputs, for example from higher visual areas, can modulate competition in primary visual cortex. Or we could investigate feature binding in the representations of compound visual objects (in cortical area IT, for example).

Maintaining state with cortical oscillations

In the last couple of years we have explored the properties of a simple model for cortical oscillations (see a brief description of the work). Our results showed the possibility of maintaining state using oscillating units that are realistic for cortex, but this possibility needs to be explored. In this workgroup we would begin by discussing the properties of the basic model, and then propose some simple ways to verify and use oscillations for maintaining state in cortex. The workgroup could go in either an analytical or an applied direction, depending on interest; this model is also suitable for implementation in neuromorphic hardware, if an engineer is interested in adapting it for their hardware.

Neural Computation in cortical circuits discussion group

Each year at Capo Caccia we have a series of highly interesting discussions and mini tutorials centred around computation in cortex, and in abstract network architectures. When can competitive interactions exist in cortex? What computational or information-processing behaviours can they implement? What other computational models are realistic for cortex? We would like to hear what you've been thinking about and working on for the last twelve months! Please come if you are working on computation in networks, or know about cortical anatomy and how that would constrain computational mechanisms, or are making experimental measurements of function in cortex, or are generally just strongly opinionated about things.

About Capo Caccia

The Capo Caccia cognitive neuromorphic engineering workshop is a two-week meeting, modelled on the successful Telluride workshop. Participants spend each morning in a series of curated discussions covering topics from cognition, through neurobiology, to neuromorphic engineering. The afternoons are reserved for self-organised projects and discussion groups, which can be proposed by any of the attendees. The flavour of the workshop is very fluid; and by virtue of the secluded location (15 minutes to the nearest town) the workshop is very intense! Senior attendees and students spend breakfast, lunch and dinner together — a great opportunity to discuss science with very knowledgeable people.

Photo map reference.