Neural Material Memory
A research programme about history-dependent responses, competing explanations and what another measurement could tell us.
The question
Neural Material Memory asks how to distinguish competing explanations of a material’s response when its previous history matters. If several models can account for the observations, what would justify preferring one? What additional measurement would actually help?
The question connects modelling to experimental design. Fitting existing observations and identifying the mechanism behind them are different tasks, and a useful investigation has to examine both.
Exhibit 1DiagramRead left to right
Return-point memory, in one picture
Some materials remember the largest load they have seen: a smaller cycle inside the big one closes exactly where it opened. Return-point memory, the Kaiser and Mullins effects and populations of simple two-state elements are the frameworks I am using to ask whether neural tissue does anything similar.
The current work
I am developing the computational part of the programme, including checks on uncertainty and on how strongly the available observations constrain an explanation. The aim is to make the comparison between models explicit and open to challenge.
An informative calculation should tell us where the evidence is strong and where different explanations remain difficult to separate. It should also clarify which assumptions carry the conclusion.
Exhibit 2NumbersEach figure links to the record that states it
Where the programme stands
From computation to experiment
The programme separates an initial computational study from a later study with controlled material experiments. The experimental stage is a planned direction. This remains research in development.
My interest is in the next question as much as the present fit: how can a measurement be chosen because it tests an explanation, and how do we recognise when the result remains inconclusive?
Exhibit 3ResourcesOpen the thing itself, not a description of it