Utilising hippocampal neuronal calcium activity in mouse CA1 for a multimodal optical brain-computer interface

The hippocampus has been proposed to integrate information from multiple sensory modalities, supporting a comprehensive “cognitive map” for both spatial and non-spatial information. Previous studies have demonstrated decoding of hippocampal spatial information in real time by recording neuronal action potentials with electrodes. However, decoding of hippocampal non-spatial information robustly in real-time has not been previously shown. Here, we utilise the advantages of widefield optical calcium imaging to construct an optical brain-computer interface (BCI) driven by calcium activity of large neuronal ensembles (∼600 neurons) to decode spatial, visual and auditory information effectively in real time. We developed a high speed end-to-end analysis workflow with advanced machine learning techniques for decoding. This methodology achieves high decoding accuracy and provides a “cognitive translation” approach that may be applied to both research and clinical applications to allow direct neural communication with animals and patients with impairment of function.