Biomedical Engineering Reference
In-Depth Information
CHAPTER 9
The Neural Code
Over the past several decades, two primary thrusts have driven neuroscience research.The first thrust has
been to understand how neurons
encode
information and communicate with other neurons. The second
thrust has been to
decode
neural messages by listening in on the chatter of neurons with extracellular
electrodes. These two thrusts are by no means separate, and many researchers work in both areas. As the
encoding and decoding of information in the brain is a rich field of study, we will only cover the most
widely used techniques here. In this chapter, we will first examine how neurons use firing rates to encode
information and then turn to the basics of how to decoded the conversation.
9.1 NEURAL ENCODING
Experiments have shown that the shapes of action potentials do not vary significantly from neuron to
neuron. It is assumed, therefore, that information is not transmitted from neuron to neuron encoded in
the shape of the action potential.What does vary significantly from neuron to neuron is the rate at which
they may fire. Furthermore, the firing rate, or frequency, of one neuron can be modulated by the synaptic
inputs to that neuron. The firing of a neuron can also affect the firing rate of the neurons to which it is
connected. A central principle of neuroscience is therefore that neural information is encoded in
firing
frequency
.
9.1.1 The Membrane as a Frequency Detector
To understand how it is possible for neurons to encode information in the frequency of firings, consider a
single current pulse applied to a membrane. In Sec. 2.3 we found that if a large stimulus was applied for a
short time, the membrane may not fully charge to threshold.When the current was turned off, it will took
take time for the membrane to discharge back to rest. Now, consider that a train of short-duration stimuli
are applied to the membrane. If the duration between stimuli is long, the membrane will have time to
fully discharge after each stimulus is turned off. If the stimulus interval time is decreased, the membrane
will not fully discharge before the next stimulus is applied. The result is that the membrane potential
will
drift
toward more depolarized potentials. This phenomenon is called
temporal integration
or
temporal
summation
. In Fig. 9.1, the stimulus duration and amplitude remained constant and only the frequency
was changed. The red plots show a low-frequency stimulus where the drift does not reach threshold. The
black plots are for a higher frequency where the drift does reach threshold. These two different behaviors
demonstrate how the frequency of stimulation is capable of modulating neural behavior at the cellular
level. A secondary way that frequency can alter neural behavior is in the timing. Even when the rate is
fast enough to induce an action potential to fire, the frequency can alter
when
the action potential fires.
For example, in Fig. 9.1 a higher stimulus frequency caused an action potential to fire earlier.
9.1.2 The Synapse as a Frequency Detector
A similar phenomenon of temporal integration can occur at the synapse but at a slower rate. For example,
if a number of action potentials reach the end of the axon terminal in rapid succession, short bursts of
neurotransmitter will be released into the cleft faster than it can be cleared. Therefore, the concentration
