Figure: (A-B) Example segment of dual neck connective and motor program recording. (A) Raw voltage traces from 32-channel microelectrode array in neck connective, (B) Raw voltage traces from differential wire pair recordings in each of the 10 major flight muscles. (C) Raster plot of example spike sorted result from an experiment, where each raster line is one spike, and each row is one identified neuron, also referred to as a unit. Neurons are ordered from top-bottom by decreasing firing rate. The start of scattered bouts of flapping activity is indicated by a vertical line. Note the units that only fire during flapping bouts, in comparison to continuously active neurons.
Because millisecond-level spike timing precision is present across the entire motor program, not just in specialized muscles, encoding of this precision must either 1) be preserved throughout the nervous system from sensory inputs to interneurons and then out to the periphery, or 2) be accomplished in the periphery through local sensory inputs to motor neurons and transformation of descending commands to a more precise timescale. To test if precision is preserved and descending information carries temporally precise information about motor output, we recorded a comprehensive motor program simultaneous with population-level microelectrode array recordings from the neck connective and high-speed videography of wing kinematics. Thus, we now have a dataset that captures (to spike-level resolution) all the motor activity that describes wing motion as well as a substantial sample of descending information in awake, flapping hawkmoths.
This dataset has been collected from 16 individuals. Given the high-throughput nature of this experiment and the additional layers of spatial information present when recording from 32 channel microelectrode arrays, we adopted automated techniques with manual curation to identify individual neurons in the neck connective. Of the recordings analyzed so far, at least 10-20 individual neurons have been identified with a high degree of confidence, representing an approximately 1-2% sampling of the information in the neck connective. The neurons detected fall into different classes of behavior, with some demonstrating continuous, spontaneous activity whereas others are only active during flapping. Analysis of this dataset is ongoing and it represents one of the first new experiments unique to the FLAP MURI.