The unitary modification rules for neural networks with excitatory and inhibitory synaptic plasticity

Comparison of Pattern Discrimination Mechanisms of Hebbian and Spatiotemporal Learning Rules in Self-Organization

The spatiotemporal learning rule (STLR) proposed based on hippocampal neurophysiological experiments is essentially different from the Hebbian learning rule (HEBLR) in terms of the self-organization mechanism. The difference is the self-organization of information from the external world by firing (HEBLR) or not firing (STLR) output neurons. Here, we describe the differences of the self-organization mechanism between the two learning rules by simulating neural network models trained on relatively similar spatiotemporal context information. Comparing the weight distributions after training, the HEBLR shows a unimodal distribution near the training vector, whereas the STLR shows a multimodal distribution. We analyzed the shape of the weight distribution in response to temporal changes in contextual information and found that the HEBLR does not change the shape of the weight distribution for time-varying spatiotemporal contextual information, whereas the STLR is sensitive to slight differences in spatiotemporal contexts and produces a multimodal distribution. These results suggest a critical difference in the dynamic change of synaptic weight distributions between the HEBLR and STLR in contextual learning. They also capture the characteristics of the pattern completion in the HEBLR and the pattern discrimination in the STLR, which adequately explain the self-organization mechanism of contextual information learning.

A possible mechanism for the effect of modifiable lateral inhibition in the striatum on the selection of conditioned reflex motor responses

A mechanism is proposed for the effects of striatal dopamine-modifiable lateral inhibition on the selection of conditioned reflex motor responses. According to this mechanism, activation of dopamine D1 (D2) receptors on strionigral (striopallidal) neurons facilitates long-term depression (potentiation) of the inhibitory inputs simultaneously with potentiation (depression) of the excitatory inputs, of sufficient strength to open NMDA channels. For " weak" excitation, insufficient to open NMDA channels, the modification rules were of the opposite sign. Activation of presynaptic D2 (D1) receptors leads to decreases (increases) in GABA release from strionigral (striopallidal) axon terminals innervating strionigral (striopallidal) cells. As a result, dopamine-modifiable lateral inhibition simultaneously increases both the potentiation (depression) of the excitatory inputs to "strongly" activated strionigral (striopallidal) neurons, increasing (decreasing) their activity, and increases the depression (potentiation) of the excitatory inputs to the "weakly" activated strionigral (striopallidal) neurons, decreasing (increasing) their activity. Subsequent reorganization of neuron activity in the cortex-basal ganglia-thalamus-cortex circuit facilitates selection of conditioned reflex motor responses by further increasing (decreasing) the activity of those motor cortex neurons which were "strongly" ("weakly") excited by the striatum in conditions of dopamine release in response to the conditioned stimulus.

A possible mechanism for the effect of neuromodulators and modifiable inhibition on long-term potentiation and depression of the excitatory inputs to hippocampal principal cells

A postsynaptic mechanism for the influences of various neuromodulators and modifiable disynaptic inhibition on long-term potentiation and depression of the excitatory inputs to granule and pyramidal neurons in the hippocampus is described. According to this mechanism, facilitation of the induction of long-term depression/potentiation at the excitatory input to the inhibitory interneuron induced by the action of a neuromodulator on a receptor bound to a G(i/0)/(Gs or G(q/11)) protein can lead to decreases/increases in GABA release, weakening/strengthening of the inhibitory action on the target cell, and improvement in the conditions for induction of long-term potentiation/depression of the excitatory input to this cell. In the absence of inhibition, the same neuromodulator, activating the same type of receptors on the target cell, would facilitate induction of long-term depression/potentiation in that cell. The resultant effect of the action of the neuromodulator on the target cell depends on the ratio of the "strengths" of the excitatory and inhibitory inputs to the cell, on the presence on the interneuron and the target cell of the same or different types of receptors sensitive to this neurumodulator, and on the concentration of the neurumodulator, because of its different affinities for the receptors through which its differently directed effects on postsynaptic processes are mediated. Predictions based on this mechanism are in agreement with known experimental data.

The involvement of dopamine in strengthening cortical signals activating NMDA receptors in the striatum (a hypothetical mechanism)

A possible mechanism is proposed for the enhancement/weakening of those cortical signals in the cortex-basal ganglia-thalamus-cortex neural network which induce/do not induce opening of NMDA channels in the spiny neurons of the striatum and which can be regarded as "strong"/"weak" in terms of this measure. The mechanism is based on the modulatory influences of dopamine on changes in the efficiency of corticostriatal inputs. In the absence of dopamine, relative increases in the intensity of "strong" ("weak") cortical signals can lead to the induction of long-term potentiation (depression) of corticostriatal synapses. In this case, because of the differently directed influences on thalamic cells of signals passing via strionigral and striopallidal cells, "strong" signals at the output of the thalamus are weakened, while "weak" signals are strengthened. Activation of dopamine D1 (D2) receptors on strionigral (striopallidal) neurons may facilitate increases in the extent of long-term potentiation/depression (decreases in the extent of long-term potentiation/depression or induction of long-term potentiation/depression). The consequence of this is that "strong" signals at the output of the thalamus can be strengthened synergistically, while "weak" signals can be weakened synergistically. Background cortical signals evoking tonic release of dopamine in the striatum can decrease strengthening because of weakening of the modulatory influence of dopamine on the modification of corticostriatal synapses.

The cortico-basal ganglia-thalamocortical circuit with synaptic plasticity. II. Mechanism of synergistic modulation of thalamic activity via the direct and indirect pathways through the basal ganglia

A possible mechanism underlying the modulatory role of dopamine, adenosine and acetylcholine in the modification of corticostriatal synapses, subsequent changes in signal transduction through the "direct" and "indirect" pathways in the basal ganglia and variations in thalamic and neocortical cell activity is proposed. According to this mechanism, simultaneous activation of dopamine D1/D2 receptors as well as inactivation of adenosine A1/A(2A) receptors or muscarinic M4/M1 receptors on striatonigral/striatopallidal inhibitory cells can promote the induction of long-term potentiation/depression in the efficacy of excitatory cortical inputs to these cells. Subsequently augmented inhibition of the activity of inhibitory neurons of the output nuclei of the basal ganglia through the "direct" pathway together with reduced disinhibition of these nuclei through the "indirect" pathway synergistically increase thalamic and neocortical cell firing. The proposed mechanism can underlie such well known effects as "excitatory" and "inhibitory" influence of dopamine on striatonigral and striatopallidal cells, respectively; the opposite action of dopamine and adenosine on these cells; antiparkinsonic effects of dopamine receptor agonists and adenosine or acetylcholine muscarinic receptor antagonists.

The cortico-basal ganglia-thalamocortical circuit with synaptic plasticity. I. Modification rules for excitatory and inhibitory synapses in the striatum

It is pointed out that Ca(2+)-dependent modification rules for NMDA-dependent (NMDA-independent) synaptic plasticity in the striatum are similar to those in the neocortex and hippocampus (cerebellum). A unitary postsynaptic mechanism of synaptic modification is proposed. It is based on the assumption that, in diverse central nervous system structures, long-term potentiation/depression (LTP/LTD) of excitatory transmission (depression/potentiation of inhibitory transmission, LTDi/LTPi) is the result of an increasing/decreasing the number of phosphorylated AMPA and NMDA (GABA(A)) receptors. According to the suggested mechanism, Ca(2+)/calmodulin-dependent protein kinase II and protein kinase C, whose activity is positively correlated with Ca(2+) enlargement, together with cAMP-dependent protein kinase A (cGMP-dependent protein kinase G, whose activity is negatively correlated with Ca(2+) rise) mainly phosphorylate ionotropic striatal receptors, if NMDA channels are opened (closed). Therefore, the positive/negative post-tetanic Ca(2+) shift in relation to a previous Ca(2+) rise must cause NMDA-dependent LTP+LTDi/LTD+LTPi or NMDA-independent LTD+LTPi/LTP+LTDi. Dopamine D(1)/D(2) or adenosine A(2A)/A(1) receptor activation must facilitate LTP+LTDi/LTD+LTPi due to an augmenting/lowering PKA activity. Activation of muscarinic M(1)/M(4) receptors must enhance LTP+LTDi/LTD+LTPi as a consequence of an increase/decrease in the activity of protein kinase C/A. The proposed mechanism is in agreement with known experimental data.

Interrelated modification of excitatory and inhibitory synapses in three-layer olivary-cerebellar neural network

The model of three-layer olivary-cerebellar neural network with modifiable excitatory and inhibitory connections between diverse elements is suggested. The same Hebbian modification rules are proposed for Purkinje cells, granule (input) cells, and deep cerebellar nuclei (output) cells. The inverse calcium-dependent modification rules for these cells and hippocampal/neocortical neurones or Golgi cells are conceivably the result of the involvement of cGMP and cAMP in postsynaptic processes. The sign of simultaneous modification of excitatory and inhibitory inputs to a cell is opposite and determined by the variations in pre- and/or postsynaptic cell activity. Modification of excitatory transmission between parallel fibers and Purkinje cells, mossy fibers and granule cells, and mossy fibers and deep cerebellar nuclei cells essentially depends on inhibition effected by stellate/basket cells, Golgi cells and Purkinje cells, respectively. The character of interrelated modifications of diverse synapses in all three layers of the network is influenced by olivary cell activity. In the absence (presence) of a signal from inferior olive, the long-term potentiation (depression) in the efficacy of a synapse between input mossy fiber and output cell can be induced. The results of the suggested model are in accordance with known experimental data.