Synaptic gradients transform object location to action

Abstract

To survive, animals must convert sensory information into appropriate behaviours1,2. Vision is a common sense for locating ethologically relevant stimuli and guiding motor responses3–5. How circuitry converts object location in retinal coordinates to movement direction in body coordinates remains largely unknown. Here we show through behaviour, physiology, anatomy and connectomics in Drosophila that visuomotor transformation occurs by conversion of topographic maps formed by the dendrites of feature-detecting visual projection neurons (VPNs)6,7 into synaptic weight gradients of VPN outputs onto central brain neurons. We demonstrate how this gradient motif transforms the anteroposterior location of a visual looming stimulus into the fly’s directional escape. Specifically, we discover that two neurons postsynaptic to a looming-responsive VPN type promote opposite takeoff directions. Opposite synaptic weight gradients onto these neurons from looming VPNs in different visual field regions convert localized looming threats into correctly oriented escapes. For a second looming-responsive VPN type, we demonstrate graded responses along the dorsoventral axis. We show that this synaptic gradient motif generalizes across all 20 primary VPN cell types and most often arises without VPN axon topography. Synaptic gradients may thus be a general mechanism for conveying spatial features of sensory information into directed motor outputs.

Organization of descending neurons in the brain of the desert locust

“…The large number of functionally characterized DBNs in the desert locust and their assignment to groups that are highly homologous in polyneopteran insects and also correspond to DBN groups in holometabolous species will greatly facilitate comparing DBN functions across insects. Major progress in understanding task sharing across populations of DBNs has recently been possible in the fly Drosophila , largely through optogenetic analysis of DBNs involved in flight control, directional escape, walking, grooming, and other behaviors (Cande et al., 2018; Dombrovski et al., 2023; Namiki et al., 2022). The detailed analysis of locust brain areas innervated by intersegmental neurons will contribute to further uncover pathways from sensory neuropils like the antennal and optic lobes to DBNs.…”

“…In walking insects, DBN activity has been shown to correlate with forward speed (Böhm & Schildberger, 1992; Cande et al., 2018; Hörner, 1992) or rotational velocity (Böhm & Schildberger, 1992; Chen et al., 2018; Hörner, 1992; Rayshubskiy, 2020; Staudacher, 2001; Staudacher & Schildberger, 1998). Some DBNs are involved in the pursuit of small targets (Milde & Strausfeld, 1986; Strausfeld & Gronenberg, 1990) or in escape jumps and escape maneuvers (Dombrovski et al., 2023; Fotowat et al., 2011; Pearson et al., 1980; Santer et al., 2008). Others trigger fast saccadic turns in flying Drosophila (Schnell et al., 2017) or escape runs in walking cockroaches (Comer et al., 1994; Stierle et al., 1994).…”

One-to-one mapping between deep network units and real neurons uncovers a visual population code for social behavior

“…5 b , tick lines) if at least 5 synaptic connections existed between an LC/LPLC neuron type and another neuron type; we chose 5 synapses to account for the fact that some neuron types had 1-2 neurons in total. Previous studies typically set a threshold of 50 synapses for a group of neuron types (e.g., a connection exists if 50 synapses are present between all LC4 neurons and all DN neurons, regardless of the specific DN cell type) [10, 41, 79]; instead, we do not group neuron types in order to have finer granularity. We identified 417 presynaptic partners and 660 postsynaptic partners (each partner being one neuron type).…”

“…The 1-to-1 network never had access to real LC responses during training (behavior only) and could only use one internal model LC unit for prediction. We targeted LC6, LC11, LC12, LC15, and LC17, each of which has neurons whose dendrites span the lobula and converge onto the same optic glomerulus (inset for LC11 neurons taken from FlyWire connectome [41] with the Neuroglancer software [42]). We image and take as the response the summed calcium dynamics within the region occupied by the glomerulus (insets, ‘stim-off’/‘stim-on’: activity before/during visual stimulus was presented).…”

“…Interestingly, all LC types responded reliably to the varying of female position (‘vary female position’, color traces) despite the facts that the optic glomeruli have weak retinotopy [11, 41] and that the calcium trace is a sum of the activity of almost all neurons for the same LC type (presumably averaging away any spatial information). This suggests that either our targeted region for calcium imaging (Fig.…”