Memory is a fundamental capacity of the human brain.  Memory binds together the past and the present, and gives us the ability to function in modern society.  Memory deficits are among the most devastating symptoms of certain brain disorders, such as epilepsy, stroke and Alzheimer’s disease.  A challenge for neuroscience in the 21st century is to understand how the brain generates memory, that is, how it encodes, stores, and retrieves information. 

The reductionist approach of recording from single neurons and single synapses has provided a wealth of data, but a unified framework for understanding how neurons act collectively to process information remains to be developed.  New technology, such as multielectrode arrays and imaging, enables us to approach such questions at the network level, and principles of biologically-realistic neural information processing are beginning to emerge.

Current Research Programme

The overall aim of our research programme is to understand how the cortex of the brain encodes, stores, and retrieves information.  Cortical neural activity displays characteristic rhythmic patterns, reflected in the human electroencephalogram as behavioural state-dependent oscillations in different frequency bands.  By organising spikes into specific temporal structures, network oscillations are conducive to spike timing-dependent synaptic plasticity, a leading candidate cellular mechanism underlying memory formation.  We aim to understand the relation between network oscillations and synaptic plasticity in the cortex, and how these temporal structures optimize encoding and retrieval of information.  To achieve this aim, our research follows three main strands, as described in the following. 

The mechanisms underlying synchronization of neuronal activity

Our research has uncovered the importance of GABAergic interneurons for the control of spike timing.  By interacting with intrinsic conductances in principal neurons, individual GABAergic interneurons are sufficient to synchronise spike discharge in pairs of pyramidal neurons, and, by inference, all of the principal neurons contacted by that same interneuron.  We discovered that the feedback circuitry of principal neurons and GABAergic interneurons in hippocampal area CA3 is sufficient to maintain a fast network oscillatory state (~40 Hz) under cholinergic activation.  The mechanism underlying this cholinergically-induced oscillation involves both excitatory and inhibitory synapses as well as intrinsic conductances in neurons, and the oscillation specifically engages, and depends upon, a subpopulation of perisomatic-targeting interneurons.  These findings have implications for our understanding of how information can be stored in and retrieved from cortical circuits. 

Synaptic plasticity during network oscillations 

Network oscillations naturally organise spike timing in key elements of the neural circuit and could thus form a basis for spike timing-dependent synaptic plasticity.  Our research has confirmed that activation of presynaptic neurons immediately before single spikes in the postsynaptic neuron is sufficient to induce input-specific synaptic potentiation in hippocampal slices prepared from rodents early in their postnatal development.  In adult animals, however, no such potentiation was seen.  Rather, there was a requirement of postsynaptic burst firing for potentiation to occur, due to increasing GABAergic inhibition with developmental maturation.  This suggests that different logical rules operate at different stages of development, and that different codes might apply during encoding and retrieval of information.  

The retrieval of information stored by a spike timing-dependent plasticity rule

Whereas much research has been directed towards understanding the rules of induction of synaptic plasticity, much less has been done on the retrieval of this information.  No theory of memory can be complete without this understanding.  We are currently investigating the relation between encoding and retrieval of information during network oscillations for optimal memory performance of cortical networks.  In the future we aim to bring these strands of research together by investigating the relation between network oscillations and synaptic plasticity.  This research programme promises to give insight into a cortical network code, i.e. how neurons collectively can represent information in a form conducive to storage and retrieval, and could also contribute to the understanding of disease processes affecting cortical processing, such as epilepsy, schizophrenia and Alzheimer’s disease.

Further information can be found at: http://noggin.physiol.ox.ac.uk/

Ole Paulsen