Neural Circuits Lecture 18

Lecture 17


Neural Circuits

Modelling of Neural Circuits: Examples from the pyloric CPG of lobsters
nm
What is Modelling?
 Here’s a suggestion of a definition:
 Modelling (in biology) is making a representation of a
biological system with the intent of understanding it (better).

Different types of models
 Physical models
 Model of a skeleton
 Electronic circuit emulating a neuron
 Theoretical models
 Conceptual models (“box and arrow models”)
 Mathematical models
 Computer (numerical) models
 Biological models (systems)
 “The rat brain is a model system for the human brain”

Why do we need models?
 When we have enough data about the brain, won’t we understand how it
works?
 Models force us to make assumptions explicit
 We can only get so far with hypotheses expressed in intuitive
forms (e.g. ‘Visual experience affects visual development’)
 Enables many “virtual” experiments to be done, can pinpoint the one that
is most crucial
 Experiments can pinpoint holes in our understanding
 Can lead to unexpected predictions
 Often much quicker/easier to try out ideas
 Can guide the design of potential experiments
 Can help testing the self-consistency of a set of assumptions and
hypotheses
 However, Modelling is not actual hypothesis testing
 Can’t explain everything with a model
 Just because the model works doesn’t mean the brain will

What makes a good model?
 Model assumptions are consistent with all established facts
 The model itself is self-consistent
 Reproduces known data
 The model is stable (unless the system isn’t either)
 The model is structurally stable
 The model can make non-trivial predictions
 The model makes correct predictions