Structure-Based Models of NO Diffusion in the Nervous System - Computational Neuroscience: A Comprehensive Approach - page 122

B

A

200

8

150

7

100

6

50

Synth vol.
Whole vol

5

1 2 3 4 5 1 2 3 4 5

Diameter ( µ m) Diameter ( µ m)

C D

0.6

1.2

0.5

1

0.4

0.3

0.8

0.2

0.6

0.1

1 2 3 4 5 0 50 100 150 200 250

Diameter ( µ m) Distance ( µ m)

Figure 4.9
Mean results for plexuses composed entirely of fibres of diameters d = 1 , 2 , 3 , 4or5µ m in a
100 × 100 × 100µ m 3 synthesising volume for 1 second of NO synthesis. Results averaged over
30 random plexuses grown generated using the growth algorithm detailed in [35]. A. Mean
delay until interaction against diameter of plexus fibres. B. Mean volumes over threshold
both in total and inside the synthesising volume against diameter of plexus fibres. C. Mean
maximum concentration obtained against diameter of plexus fibres. D. Mean concentrations
seen in a 1d line through the centre of either thin (1

µ m ) or thick (5 µ m ) plexuses.

4.4 Exploring functional roles with more abstract

models

The computational models presented above require huge numbers of iterated calcu-
lations and inevitably place heavy demands on processing resources. Hence it is not
yet feasible to build models of whole neuronal networks at that level of detail and
run them in real time, or anything even vaguely approaching it. Therefore, in paral-
lel with the detailed modelling work, we have developed a class of computationally
fast artificial neural networks (ANNs) that incorporate a more abstract model of sig-
nalling by diffusing neuromodulators.

Such networks have been used as artificial