Correlated Neuronal Activity: High- and Low-Level Views - Computational Neuroscience: A Comprehensive Approach - page 341

are stimulus-dependent, just like sensory-evoked firing rates. The studies discussed
below suggest a different alternative in which correlations change rapidly as func-
tions of internal events and may regulate the flow of neural information, rather than
its meaning [68].

12.2.2 Information flow

The regulation of information flow is illustrated by the following result [70]. When
intracortical microstimulation is applied during performance of a visiual-motion dis-
crimination task, the subject's response is artificially biased, but the bias depends
strongly on the time at which the microinjected current is delivered relative to stim-
ulus onset. Microstimulation has a robust effect if applied during presentation of the
visual stimulus, but it has no effect if applied slightly earlier or slightly later than
the natural stimulus [70]. This suggests that even a simple task is executed accord-
ing to an internal schedule, such that the information provided by sensory neurons
is effectively transmitted only during a certain time window. How does this inter-
nal schedule work? One possibility is that changes in correlations are involved [68].
This is suggested by a number of recent experiments in which correlations were seen
to vary independently of stimulation conditions. To work around the usual prob-
lems with stimulus-linked correlations, investigators have studied correlated activity
in paradigms where, across trials, stimulation conditions remain essentially constant
and the most significant changes occur in the internal state of a subject.
Riehle and colleagues trained monkeys to perform a simple delayed-response task
where two cues were presented sequentially [66]. The first cue indicated a target
position and instructed the animal to get ready, while the second cue gave the go
signal for the requested hand movement. Crucially, the go signal could appear 600,
900, 1200 or 1500 ms after the first cue, and this varied randomly from trial to
trial. Neurons recorded in primary motor cortex increased their synchrony around
the time of the actual sensory stimulus or around the time when the animal expected
the go signal but it did not appear [66]. The latter case is the most striking, because
there the firing rates did not change and neither did the stimulus; the synchronization
depended exclusively on the internal state of the monkey.
Fries and colleagues [35] used attention rather than expectation to investigate the
synchrony of visual neurons in area V4. They used conditions under which firing
rates varied minimally, taking advantage of the finding that, although attention may
have a strong effect on the firing rates evoked by visual stimuli, this modulation
is minimized at high contrast [65]. Monkeys were trained to fixate on a central
spot and to attend to either of two stimuli presented simultaneously and at the same
eccentricity. One of the stimuli fell inside the receptive field of a neuron whose
activity was recorded. Thus the responses to the same stimulus could be compared
in two conditions, with attention inside or outside the neuron's receptive field. At the
same time, the local field potential (LFP) was recorded from a nearby electrode. The
LFP measures the electric field caused by transmembrane currents flowing near the
electrode, so it gives an indication of local average activity [38]. The correlation that
was studied in these experiments [35] was that between the LFP and the recorded