“…Previous work indicates that subjects can learn to control neuroprosthetic devices using single cells or bulk electrophysiological signals (Fetz, 1969;Bakay and Kennedy, 1998;Nicolelis, 2001;Serruya et al, 2002;Carmena et al, 2003;Weiskopf et al, 2003;Sitaram et al, 2007;Koralek et al, 2012;Hochberg et al, 2012;Collinger et al, 2013;Clancy et al, 2014;Sadtler et al, 2014;Prsa et al, 2017;Sitaram et al, 2017;Trautmann et al, 2019), but this is the first work, to our knowledge, to employ control using imaged population calcium signals. This technique allowed us to monitor much of the dorsal cortical network as animals learned neuroprosthetic control, whereas previous BMI work has been limited to recording from neighbouring neurons (Koralek et al, 2012;Clancy et al, 2014;Sadtler et al, 2014;Prsa et al, 2017). Using population signals, rather than individual neurons, to manipulate neuroprosthetic devices might afford more stable and minimally invasive control, robust to losing signals from single control cells.…”