Back to Main TMM page
The NIH BRAIN Initiative would like to know your thoughts on how to integrate theories to understand the brain!
Summary discussion and live document on future development of theoretical neuroscience
During the BRAIN Initiative PI Meeting held virtually from June 1 to June 2, 2020 several topics were discussed in a virtual Networking Lounge. The question formulated in this chat was, What are the roadblocks in generating new theories and integrating competing theories in neuroscience?
Here is the summary of the conversation
This is the transcript
We have posted copied the Summary file into a editable Google Docs for all of you to expand and contribute to the discussion.
A similar recent effort
Fidel Santamaria co-organized a workshop on Present and Future Frameworks of Theoretical Neuroscience (supported by NSF), which addresses similar issues discussed at the TMM PI meeting. Here is the arXiv summary we published.
NSF Workshop Report
A paper addressing structure of neuronal theories was produced by one of the workgroups who attended this workshop.
This review article provides a high level review of theoretical and empirical work suggesting a theoretical framework for how the brain might think. The basic idea is that populations of neurons represent information over continuous dimensions such as time and space, but also other more abstract dimensions. The theory suggests that the brain uses the Laplace transform to build and manipulate these representations.
Howard, M.W. and Hasselmo, M.E. (submitted). Cognitive computation using neural representations of time and space in the Laplace domain.
This paper shows evidence for the real Laplace transform of time in entorhinal cortex, confirming a prediction of the theory going back more than a decade.
Bright, I.M.*, Meister, M.L.R.*, Cruzado, N.A., Tiganj, Z., Buffalo, E.A.*, and Howard, M.W.* (In press).
A temporal record of the past with a spectrum of time constants in the monkey entorhinal cortex. PNAS
- A Python package for data-driven validation of computational models (or all types, including dynamical systems): SciUnit
- A SciUnit-based Python package for data-driven validation of single-neuron models: NeuronUnit
- Intro to the model validation tools above, including teaser for web application for visualizing and scheduling model validation experiments.
- An alternative to ModelsDB for neuron, network, and channel models: NeuroML-DB. All models actually run, and standardized simulation experiments have been run against them (and are shown there).
- A database of (mostly slice, mostly rodent) intracellular slice physiology data: NeuroElectro.