One of the greatest, relatively underappreciated,
discoveries in all of science was the discovery of the nerve impulse in the
1930s by the British Lord Adrian. Adrian did win a Nobel Prize for his
discovery in 1932, but scholars underestimated its implications, which go
beyond the fact that four later Nobel Prizes were awarded for work based on
Adrian’s discovery. This included discovery of sodium and potassium ionic flux
during impulses, the role of impulses in releasing neurotransmitters, and the
role of membrane ion channels in impulse generation and second messenger
cascades.
Like many discoveries in science, this one could not have
been made without technological advance. In this case, the essential advance
was the development of the capillary electrometer, which enabled detection of
very small electrical pulses on the order of one millisecond duration. This
instrumentation was crude and far inferior to later advances such as the oscilloscope
and computer screens. Before Adrian’s use of the electrometer, scientists
generally knew that peripheral nerves generated some kind of electrical signal,
but nothing was known about the nature of the signal in individual neurons.
Nerves contain fibers from hundreds of neurons that produce
a summed, relatively long duration and large wave that spreads down the nerve.
No one knew how the individual nerve fibers contributed to this compound
signal. Adrian answered this question by tedious microdissection of nerves into
their individual fibers and recording stimulus-evoked responses in a single
fiber. What Adrian saw was that the response was a series of voltage pulses, each
about one millisecond long, all of the same amplitude in a given fiber. Decades later, development of microelectrodes
enabled confirmation of Adrian’s discovery in neurons in the brain.
Fig. 1. Train of nerve impulses from a single neuron over
2.5 seconds, as recorded with extracellular electrodes. Amplitude calibration =
0.5 millivolts. The thick baseline is electronic noise, in which the spikes are
embedded. The signal-to-noise ratio is vastly improved with modern electronics
and intracellular recording. From Fromm and Bond, 1967, Electroenceph. clin.
Neuro. 22, 159.
This provided the evidence of the basic similarity and
difference between brains and the later development of computers. Both
computers and brains convert the real world into representations. In computers,
information is coded, in the form of 1s and 0s, and as nerve impulses in
brains. Both computers and brains distribute and process this represented information,
and can store it as memories. However, because brains are biological and use
impulses to represent information, they can change their circuitry and can
self-program. Unlike computers, brains also have will, including a likely
degree of free will.
Brains have conspicuous functional states, ranging from
intense conscious concentration to drowsiness, to sleep, to coma, to death. Neuronal
electrical activity correlates in a systematic way with these state changes. The
most conspicuous of these activity measures exist in terms of nerve impulse
firing and the extracellular ionic currents they create at synapses, known as
field potentials. As these field potentials reach the scalp, they produce the
signal we call an electroencephalogram. Field potentials are technologically
easier to record than individual nerve impulses, but more ambiguous to interpret
because of the spatial summation of voltages from hundreds of heterogeneous
neurons.
The original nerve impulse findings were that the rate of
impulse firing governed the impact on neuronal targets, whether they be muscle
or other neurons. Various labs, including my own, in the 1980s discovered that
the intervals between impulses also contained their own kind of information.
For example, my lab reported that some neurons contained statistically
significant serial ordering of impulse intervals in a neuron’s impulse stream.
The intervals, at least in higher-level brain areas, are not random. They are
serially dependent, as if they contained a message. If you are familiar with
Markov transition probability, you can understand our finding that serial
dependences exist in as many as five successive intervals (Sherry et al. 1982).
This led us to suggest “byte processing” as a basic feature of neuronal
information processing. This view has not caught on, and most people still seem
to think that firing rate is the basic information code, despite the
well-established temporal summation that occurs as impulses arrive at synapses.
Bernard Katz demonstrated temporal summation of impulse effects in
neuromuscular junctions in 1951 and later J.P. Segundo and colleagues confirmed
it in neuronal synapses (Segundo et al., 1963).
It should not be
surprising that there are serial dependencies in impulse intervals. For
example, intracellular recording of postsynaptic potentials revealed that the
polarization change caused by a single impulse input decays in a few
millisecond. However, a succession of closely spaced impulse inputs allows the polarization
changes to summate.
These days, the emphasis needs to be put on impulse activity
in defined circuitry. All neurons are linked in one or more circuits, and the
impulse train in any one neuron is only a small part of the over-all circuit
activity. The function of any given circuit depends on the circuit impulse
pattern (CIP) of the whole circuit. Researchers have developed microelectrodes
that allow recording of impulse trains from single neurons, but the problem is
in implanting a series of electrodes so that each one monitors the activity of
a selected neuron in a defined.
I think that research should focus on CIPs and the phase
relationships of electrical activity among cortical circuits, both within and
among cortical columns (Klemm, 2011). Nerve impulses have to be at the heart of
consciousness, inasmuch as impulses contain the brain’s representation of
information and create the synaptic field potentials.
We know from monitoring known anatomical pathways for
specific sensations that the brain creates a CIP representation of the stimuli.
As long as the CIPs remain active, the representation of sensation or neural
processing is intact and may even be accessible to consciousness. However, if
something disrupts ongoing CIPs to create a different set of CIPs, as for
example would happen with a different stimulus, then the original
representation disappears. If the original CIPs persist long enough, a memory
could form, but otherwise the information would be lost. The implication for
memory formation is that the immediate period after learning must be protected
from new inputs to keep the CIP representation of the learning intact long
enough to form a more lasting memory.
Much current research shows that conscious awareness
correlates with the degree of synchrony and time-locking of CIPs in various
regions and within regions of cortex. The evidence comes from
electroencephalographic monitoring of the oscillating field potentials in a
given area. These are voltage waves that occur in multiple frequency bands.
Phase relationships of voltage waves from different circuits surely reflect the
timing of the impulse discharges that create those fields. I summarized the
animal research evidence for this view in my first book, some 50 years ago
(Klemm, 1969). Depending on the nature of stimulus and mental state, these
oscillations of various circuits may jitter with respect to each other or
become time locked. The functional consequence of synchrony has to be
substantial, and many others and I suggest that this is a fundamental aspect of
consciousness. The correlation between frequency coherences and states of consciousness
is clear. Frequency coherence reflects a “binding” of neurons into linked and
shared electrochemical activity, but how this relates to conscious awareness
will require a next great discovery in science.
Sources:
Klemm, W. R. (1969). Animal Electroencephalography. New
York: Academic Press.
Klemm, W. R. (2011). Atoms of Mind. The “Ghost in the
Machine” Materializes. New York: Springer.
Segundo, J. P., et al. (1963). Sensitivity of neurons in Aplysia to temporal pattern of arriving
impulses. J. Exp. Biol. 40: 643-667.
Sherry, C. J., Barrow, D. L., and Klemm,
W. R. 1982. Serial dependencies and Markov processes of neuronal interspike
intervals from rat cerebellum. Brain Res. Bull. 8: 163‑169.
