RAJAN, KANAKA Multi-region Network of Networks Recurrent Neural Network Models of Adaptive and Maladaptive Learning EB028166
We in the Rajanlab at Mount Sinai design neural network models constrained by experimental data, and reverse engineer them to figure out how brain circuits function in health and disease. Here are two scalable, flexible, and robust tools we have developed, through our TMM grant (R01EB028166-01), that will find wide adoption across to the broader neuroscience community, particularly in a few U19 consortia. We are already collaborating fruitfully with the BRAIN_COGS consortium at Princeton University and would be delighted to share our collective findings at the upcoming meeting. That said, we are always looking for opportunities to collaborate with experimental labs and are particularly keen to get involved at the ground-floor of such collaborations, i.e., having a key role in designing experiments, not only leveraging existing data.
1) The first tool we have developed is named Current-based Decomposition or CURBD for short. This powerful new theory-based framework is intended for “in-vivo tract tracing” from multi-regional neural activity collected experimentally. CURBD employs recurrent neural networks (RNNs) directly constrained, from the outset, by time series measurements acquired experimentally, such as Ca2+ imaging or electrophysiological data. Once trained, these data-constrained RNNs let us infer matrices quantifying the interactions between all pairs of modeled units. Such model-derived “directed interaction matrices” can then be used to separately compute excitatory and inhibitory input currents that drive a given neuron from all other neurons. Therefore different current sources can be de-mixed – either within the same region or from other regions, potentially brain-wide – which collectively give rise to the population dynamics observed experimentally. Source de-mixed currents obtained through CURBD allow an unprecedented view into multi-region mechanisms inaccessible from measurements alone. We have applied this method successfully to several types of neural data from our experimental collaborators, e.g., zebrafish (Deisseroth lab, Stanford), mice (Harvey lab, Harvard), monkeys (Rudebeck lab, Sinai), and humans (Rutishauser lab, Cedars Sinai), where we have discovered both directed interactions brain wide and inter-area currents during different types of behaviors. With this powerful framework based on data-constrained multi-region RNNs and CURrent Based Decomposition (CURBD), we ask if there are conserved multi-region mechanisms across different species, as well as identify key divergences.
2) We have additionally developed a second data-inspired recurrent neural network (RNN)-based tool, termed TRAKR. This tool enables us to detect state transitions from time series data. Using TRAKR, we can identify neural state transitions that are predictive of animal behavior under different experimental and experiential conditions. For example, we can pick up behavioral, neural, and circuit mechanistic changes in decision making, those that occur during shifts in attention or task engagement in multi-task paradigms, those accompanying learning of such tasks in the lab, etc. The current implementation of TRAKR works with time series data such as behavioral monitoring from pose detection methodologies, EEGs, ECoGs, LFPs, spiking neural data, and calcium fluorescence signals. This is currently a human-in-the-loop method that detects state transitions after correlation with behavior, so behavioral recordings are helpful but not necessary. There is no threshold for sampling frequency, but larger datasets are desired since RNNs are more robust when supplied with greater amounts of data. This tool is agnostic to the species, brain areas, or cell types. It is similarly oblivious to modality and can work equally well with large amounts of EEG, LFP, and even imaging time series.