An inducing algorithm for LTP in hippocampal CA1 neurons studied by temporal pattern stimulation

To investigate the effect of the time structure of input spike trains for CA1 neurons in eliciting LTP, the authors examined the relationship between statistical properties (mean rate, serial correlation coefficient) of stimulus sequences and the induction of LTP. The statistical stimuli were Markov stimuli with different second order statistics (type 1 is positive correlations between successive inter-stimulus intervals, type 2 is negative, and type 3 is independent) but with identical mean rate. The magnitude of LTP induced by these stimuli showed clear order relationships, type 3>type 1≫control>type 2. From the experimental data, a dynamical learning rule in CA1 neural networks was derived that extracts the temporal information of input stimuli and transforms it into the weight space of synaptic connection in CA1 hippocampal networks