The Role of Correlated Synaptic Activity in Neural Integration

This page describes the activities undertaken as part of a grant with the above title, funded by EPSRC (GR/R12350/01).

Abstract. The abstract of the research proposal is available here.

Background. In this project we are interested in how spatial temporal interactions within populations of synaptic inputs shape the output discharge of a neurone. We have investigated this issue in detailed compartmental models of motoneurones and CA3 pyramidal neurones.  We characterise spatial temporal interactions by studying the effects of introducing spatial and temporal correlation into a sub set of the total synaptic input. Spatial correlation refers to clustering of inputs in certain locations, temporal correlation refers to the timing of individual spikes in the spike trains driving the synaptic inputs.

Results
First we establish the basic pattern of spatial temporal interactions within a neurone subjected to large scale synaptic input which lacks spatial or temporal correlation, i.e. synaptic inputs are uniformly distributed over the dendritic tree and all inputs are activated by random spike trains.
Local feedback in dendritic currents following individual inputs
Transfer function analysis of membrane potential fluctuations during large scale input


Local feedback in dendritic currents following individual inputs
One phenomenon we have observed is a reversal of the axial current, ia, in  dendrites following synaptic activation. This process is illustrated in the figure below. The model used was a 129 compartment model of a motoneurone, consisting of a soma and 12 tapered dendrites, with a single synaptic conductance activated in the centre of one dendrite.
Axial current reversal during synaptic input

Axial current reversal following synaptic input.
This figure illustrates the axial membrane current, ia, between a dendritic compartment and its proximal (left), and distal neighbours (right). Note the reversal of axial current after 0.9 ms, which only occurs in the distal current. The solid lines indicate the current in response to a single synaptic input applied in isolation, the dashed lines are the current due to a single input in the presence of large scale background activation of the cell.

The EPSP observed at the soma for this single input has the classical form expected of an EPSP activated from rest (Rise time 0.85 ms; Time to peak 1.98 ms; Half width 9 ms). The current reversal occurs during the initial rising phase of the EPSP. The current reversal only occurs in the distal direction. We believe this feedback results, in part, from the local geometry of  tapered dendrites. It is present both for single inputs in isolation (solid lines) and single inputs in the presence of large scale background activation (dashed lines; total of 996 inputs active). This process of local feedback may have consequences for our understanding of synaptic integration, since it results in multiple copies or “imprints” of synaptic events propagating towards the soma, on a very short time scale.

Transfer function analysis of membrane potential fluctuations during large scale input
We use a statistical signal processing framework to analyse the relationship between membrane potential fluctuations in different parts of the dendritic tree during large scale synaptic input, using the same notoneurone model as above. The figure below shows the axial membrane current and a time domain transfer function analysis of the membrane potential fluctuations. The synaptic input is uniformly distributed and all 996 inputs are activated by random spike trains, each with a mean rate of 32 spikes/sec.

Analysis of membrane potential fluctuations  and axial current during large scale synaptic input. (Left) Section of 100ms duration showing axial current between one dendritic compartment and its proximal neighbour during large scale background activation of the neurone. (Right) Estimate of time domain transfer function between the membrane potential fluctuations at the mid point of one dendrite and those at the soma for a 100 s record during large scale synaptic input.

The axial current trace, which fluctuates about zero,  indicates a continual two way flow of membrane current during large scale synaptic input. The axial current reversal mechanism described above will contribute to this process, as well as moment to moment fluctuations in the timing of individual synaptic inputs. This continual reversal of the axial current is reflected in the symmetrical form of the transfer function between the membrane potential fluctuations at the soma and those in the centre compartment of one dendrite. The membrane potential fluctuations have a broad dependence over ±10 ms. How do we interpret this transfer function? For the case of uniformly distributed random synaptic input it provides an indicator of the time scale of the dependence of the membrane potential fluctuations at the soma with those at one location in the dendrites.

Please note - this page is still under construction. Further results will be added

Further details
If you would like further details about the above work, the following publications provide details. Further publications will be added as they become available.

M. Griffin &  D.M. Halliday (2003) "On the role of dendritic feedback in synaptic integration", British Neurosci. Assoc. Abstr., vol17, p151, 2003.
Abstract is available here. Poster presentation to the BNA meeting is available in three pages (jpg): page1   page2   page 3.

M. Griffin &  D.M. Halliday (2003) "Axial current reversal promotes synchronous correlation between dendritic membrane potentials during large scale synaptic input." To be presented at the Computational Neuroscience meeting CNS03.
Pre-print (PDF) is available here. 


This work was supported by (GR/R12350/01). 

Return to Introductory Research Page

Home Page

Last Modified 05 June 2003