neural coding of behavior: evidence for precise timing of spikes ?

MGLinWS mglinws at
Thu Dec 14 15:07:27 EST 1995

sparks at (David Sparks) wrote:

:But consider motoneurons in the abducens nucleus. They have
:firing rates that are proportional to the position of the eye in the
:orbit and when the eye is in a stable position, the interspike
:intervals are remarkably constant. 

:The point of this, is that it does seem possible for neurons to maintain 
:firing rates with regular intervals. Consider too the discharge of
:cells in respiratory circuit. Does anyone have an estimate of variability
:interspike interval in these or other pacemaker cells?

Indeed, neurons in principle relay nuclei, such as the abducens, may have 
more regular patterns of spike activity than neurons in higher-order
such as the cortex or basal ganglia.  It is unlikely that all neurons in
the CNS would use an "equivalent code" and it may even "make sense" for
neural circuits to fire in regular patterns to do their thing.

I am not familiar with structures like the abducens.  How reliable are
interspike intervals 
in places like the abducens during behavior?  What is the range of the ISI
of variation there?  Has anyone attempted to relate the patterns of
activity in relay 
nuclei with behavioral performance in awake, behaving animals?  If so,
what aspects of
spike activity vary with the animal's performance?

Also, is there anything special about the membrane props or local circuits
in places like
the abducens?

For example, I suspect that the variablity I see in the striatum, e.g.,
from what 
appears to be medium spiny neurons, can be explained by the organization
of synaptic
contacts from corticostriatal cells on the spiny neurons (see anything by
Wilson for a review).  There appears to be a massive convergence of
from cortical neurons onto the dendrites of individual medium spiny cells
like 5000 to 15,000 corticostriatal neurons terminating on each spiny
cell).  It 
seems that a large, synchronous convergence of inputs from different
cortical neurons 
is required to elicit spikes in the spiny cells, due to the anomolous
in the spiny cells' dendrites.  Such a massive convergence of inputs is
bound to give
a variety of patterns of action potentials from the spiny cells on
occasion to occasion.
And it possible that the inputs might not be tightly correlated in time on
occasions.  Therefore, it is no surprise that these cells are so variable
and fail to
fire at all some of the time.

What is surprising is that no one has yet looked at whether variations in
performance are associated with variations in the temporal organization of
activity in the corticostriatal system.  Most physiologists working on the
basal ganglia 
have reported their data in terms of spike rate alone.  Now that I have
looked myself, 
I know that temporal variations do occur, and also that they do not
involve variations 
in precisely timed patterns of interspike intervals.  Moreover, temporal
features do
better than overall rates at "predicting" the animal's behavior from trial
to trial.
I now suspect that a good bit of what has been reported as evidence for
the role
of the basal ganglia in the control of behavior was actually only the "tip
of the iceberg"!
Armed with a bag of tricks for analyzing temporal variation in single
neurons and spatio-
temporal variations across ensembles of neurons, we can now begin to look
more realistically
at how neural populations really behave.

Mark Laubach
Dept. of Physiology & Pharmacology
Bowman Gray School of Medicine
Wake Forest University
Winston-Salem, NC 27157
laubach at

More information about the Neur-sci mailing list