I agree and disagree. I think neural dynamics ( and kinetics) can be understood at different levels and some do not benefit from pure mathematics but still can benefit from modelling which is not pure mathematics but a form of its application). We cannot in practice use mathematics to predict what we can measure in reality but must use modelling because the math is simply intractable. I don't think quantum computing is obviously going to doom the role of modelling but it may prove to have engineering value in probabilistic modelling. Same for brain computation. Models can provide explanations and they are not math (unless you want to say they are.) The analogy is the difference between General Relativity and actually predictions from General Relativity.

Here, at one level, is how your brain computes based on natural language evidence.

https://medium.com/liecatcher/how-your-brain-computes-41ebe7428ff9