Backward projections in the cerebral cortex: Implications for memory storage

Cortical areas are characterized by forward and backward connections betwee n adjacent cortical areas in a processing stream. Within each area there ar e recurrent collateral connections between the pyramidal cells. We analyze the properties of this architecture for memory storage and processing. Hebb -like synaptic modifiability in the connections and attractor states are in corporated. We show the following: (1) The number of memories that can be s tored in the connected modules is of the same order of magnitude as the num ber that can be stored in any one module using the recurrent collateral con nections, and is proportional to the number of effective connections per ne uron. (2) Cooperation between modules leads to a small increase in memory c apacity. (3) Cooperation can also help retrieval in a module that is cued w ith a noisy or incomplete pattern. (4) If the connection strength between m odules is strong, then global memory states that reflect the pairs of patte rns on which the modules were trained together are found. (5) If the interm odule connection strengths are weaker, then separate, local memory states c an exist in each module. (6) The boundaries between the global and local re trieval states, and the nonretrieval state, are delimited. All of these pro perties are analyzed quantitatively with the techniques of statistical phys ics.