24.2 Outstanding Problems

341

Considering that the human brain contains about10 Superscript 1110¹¹ neurons and that each neuron

may have tens of thousands of synapses, the enormous complexity of the nervous

system is evident. ²

“Both the afferent [bringing information from distant receptors] and efferent

[sending impulses to distant organs such as muscle] flow of information in the ner-

vous system is coded in terms of the identity of the activated nerve fibres and the

frequency of the impulses transmitted in them”. ³ Frustratingly, the precise infor-

mation that is conveyed between neurons remains unknown. Adrian established the

“all or nothing” nature of the nervous impulse, and also showed that the afferent

effect depends on the temporal pattern of impulses in the incoming neuron. Follow-

ing receipt of inputs, via synapses, from other neurons, possible responses include

firing a single pulse, firing a rapid sequence of pulses, switching between pulses,

and nothing (“silence”). ⁴ The choice of response depends on the internal state of

the neuron, which may also be influenced by chemicals (hormones) in its immediate

environment. Neural computation can be perceived on at least two levels, the opera-

tions actually carried out by the individual neurons, and the organization (inherently

recursive) of the operations according to the structure of the neural interconnexions. ⁵

Drawing on analogies with spin glasses, the structure of which is arranged to

minimize energy, Hopfield (1982) proposed that the firing of neurons in networks

also minimizes the energy of the network, through a process somewhat resembling

epigenetic development (Fig. 14.5).

We see that the neural network is more complex than the gene and other networks

discussed in Chap. 23, not least because the network nodes are themselves complex

entities (they could themselves be networks). Given this complexity, it is not sur-

prising that even the “simple” neural network of C. elegans’ 302 neurons and 6000

synapses is not yet fully understood, although the “wiring diagram” (connectome)

has been mapped out. ⁶

24.2

Outstanding Problems

The 11 Churchland questions are pertinent:

1. How is information coded in neural action?

2. How are memories stored and retrieved?

3. What does brain baseline activity represent?

4. How do brains simulate the future?

2 The human brain is by far the most complex nervous system. Famously, C. elegans has precisely

302 neurons. A mollusc has of the order of10 Superscript 410⁴, an insect of the order of10 Superscript 510⁵.

3 Sommerhoff (1974), p. 135; see also Panzeri et al. (2002).

4 For details of the physicochemical mechanism of transmission of signals along the nerve fibres,

see, for example, Markin et al. (1987).

5 See von Foerster (1973) for more discussion.

6 See Bargmann and Marder (2013) for a historical perspective.