Closed-loop experiments and brain machine interfaces with multiphoton microscopy

Abstract

In the field of neuroscience, the importance of constructing closed-loop experimental systems has increased in conjunction with technological advances in measuring and controlling neural activity in live animals. We provide an overview of recent technological advances in the field, focusing on closed-loop experimental systems where multiphoton microscopy-the only method capable of recording and controlling targeted population activity of neurons at a single-cell resolution in vivo-works through real-time feedback. Specifically, we present some examples of brain machine interfaces (BMIs) using in vivo two-photon calcium imaging and discuss applications of two-photon optogenetic stimulation and adaptive optics to real-time BMIs. We also consider conditions for realizing future optical BMIs at the synaptic level, and their possible roles in understanding the computational principles of the brain.

Keywords: adaptive optics; brain machine interface; closed-loop experiments; two-photon calcium imaging; two-photon optogenetics.

Fig. 1

Natural and artificial closed-loops. (a) Three learning rules and their time scales. Unsupervised learning via STDP (spike-timing-dependent plasticity), supervised learning via sensory-motor feedback, and reinforcement learning via dopamine-dependent plasticity have distinct feedback times with about an order of magnitude difference. (b) BMI using a combination of two-photon imaging (upper panel) and two-photon photostimulation (lower panel). It is thought that real-time processing within 10 ms, 100 ms, or 1 s will allow us to see the recovery and enhancement of different levels of brain functions. (c) Development of 2pBMI through a combination of two-photon imaging (green) and two-photon photostimulation (red). The horizontal axis (green) is the number of total recorded neurons during the 2pBMI experiment. The filled green is based on Ref. , but other 2pBMI studies are not much different. The horizontal axis (red) is the number of neurons per second of the two-photon photostimulation. The vertical axis shows the spatial resolution. The open circle is not the one used for 2pBMI, but is that of the experiment that can be considered the current state-of-the-art. Each arrow indicates the possible path of development that has been or will be considered.

Fig. 2

2p-BMI: (a) head-fixed mice trained to perform a lever-pull task were retrained to perform 2pSNOC (two-photon single neuron operant conditioning). The imaging area was M1 or M2. (b) ROI analysis was performed in real time. The mouse was rewarded by the activity of a target neuron, which leads to elevation of activity of the target neuron. (c) Densely expressed G-CaMP7 and sparsely expressed ChR2. (d) Photostimulation was applied 250 ms before or 2.5 s after the reward. (e) The activity of surrounding neurons was elevated when the stimulus came before the reward, and was decreased when the stimulus came after the reward, which replicated the changes in neuronal activity during 2pSNOC (adapted with permission from Ref. 6).