RESEARCH HIGHLIGHTS

NEUROSCIENCE Robots take over electrophysiology

Image-guided targeting of fluorescently labeled neurons enables automated elec- trophysiological recordings of defined neuronal cell types. Patch–clamp electrophysiology has been an invaluable technique in neuroscience; patch–clamp recordings from whole cells can provide a detailed picture of the elec- trophysiological properties of neurons. The technology involves approaching a cell with a glass micropipette, forming a seal on the membrane and breaking into the cell in order to establish an electrical continuum between the cell and the micropipette. Yet, since its invention in the seventies, this A micropipette (in red) in whole-cell patch– technique has remained labor intensive and clamp configuration during recording from an continues to require experience and skill. interneuron (in yellow) in the mouse visual cortex. Image reprinted with permission from Automated approaches would help to Annecchino et al., Elsevier Inc. increase throughput and to make the tech- nique more widely accessible. Patch–clamp mouse cortex (Suk et al., 2017). They suc- robots have been developed, but so far they cessfully established a good seal on the cell have been blind to the cell types they record in about 35% of the cases, and 60% of those from. Two groups, led by Ed Boyden from cases resulted in a whole-cell patch–clamp the Massachusetts Institute of Technology configuration. Schultz’s group targeted and by Simon Schultz from the Imperial pyramidal neurons in the primary visual College London, have independently cortex in mice and successfully achieved a reported image-guided automated patch– seal in roughly 47% of the cases, while they clamp techniques that make it possible to obtained a whole-cell patch–clamp record- record from defined cells in the mouse ing in 48% of those (Annecchino et al., brain in an automated fashion. 2017). Thus, both robots achieved a similar Some of the steps necessary for automat- overall success rate in about 20% of the tri- ed patch–clamp recording were established als, which is roughly comparable to the suc- previously, such as the approach towards cess rates of manual patch–clamp recording the target cell or the patching of random (although success depends on the skill level cells. But a key technical hurdle that had to of the experimenter). be overcome for combining the individual These image-guided patch–clamp robots achievements was tissue displacement dur- are promising developments. However, ing the approach of the micropipette. Both they have so far only been applied in groups developed similar algorithms to proof-of-principle experiments involving compensate for tissue motion, and these a limited number of trials. Their broader algorithms work by acquiring the posi- adoption will depend on their robustness tions of the micropipette tip and the target and also on further developments, such cell using two-photon imagery. Once the as the automated loading of fresh micro- pipette tip reaches the vicinity of the cell, pipettes. Furthermore, experiments that the approach path is updated in order to require recording from two or more cells at compensate for tissue displacement caused the same time would benefit if the robotic by the pipette insertion. Thus, a closed-loop capabilities could be multiplexed. feedback system keeps the approaching Nina Vogt micropipette on track. The following steps RESEARCH PAPERS of seal formation and breaking in are simi- Suk, H.-J. et al. Closed-loop real-time imaging lar to already established approaches that enables fully automated cell-targeted patch-clamp neural recording in vivo. Neuron 95, 1037–1047.e10 blindly targeted cells. (2017). Boyden’s group demonstrated the per- Annecchino, L.A. et al. Robotic automation of in formance of their robot by recording from vivo two-photon targeted whole-cell patch-clamp ­interneurons and pyramidal neurons in the electrophysiology. Neuron 95, 1048–1055.e3 (2017).

1036 | VOL.14 NO.11 | NOVEMBER 2017 | NATURE METHODS