A neural circuit for olfactory motion detection

Problem Statement
Movement is a defining feature of dynamic environments, and sensory systems across species have evolved mechanisms to extract motion direction. While visual motion detection is well-characterized as a circuit-level computation involving the comparison of signals across space and time, the neural mechanisms underlying olfactory motion detection remain unknown. Walking Drosophila utilize bilateral antennal sensing to detect the direction and speed of odor motion independently of wind sensing, yet the specific neurons and circuitry that compute this directional information from bilateral odor inputs have not been identified.

Methodology
The authors employed a multimodal approach combining in vivo two-photon calcium imaging, optogenetics, connectomics, pharmacology, and behavioral assays in adult female Drosophila melanogaster.

• Stimuli: The study utilized moving odorized ribbons (apple cider vinegar) and optogenetic "moving bars" (sequential activation of olfactory receptor neurons, ORNs, in the two antennae) to simulate odor motion. They also employed correlated noise stimuli, naturalistic odor plume waveforms, and drifting sinusoidal patterns to probe the temporal and statistical properties of the motion detector.
• Imaging: Calcium imaging (GCaMP6f) was performed in the lateral horn (LH), antennal lobe projection neuron (ALPN) terminals, and specific glomeruli (including DA1) to measure neural responses to ipsi-first (odor arriving at the ipsilateral antenna first) versus contra-first motion.
• Circuit Dissection: The authors used GABA receptor antagonists (picrotoxin and CGP 54626) to test the role of inhibition. Connectomic analysis of the FlyWire dataset was used to identify local neurons (LNs) with specific connectivity patterns (contralateral-biased input and strong output to ALPNs).
• Behavioral Assays: Freely walking flies were tested in an arena with optogenetic moving bars. The study utilized temperature-sensitive shibire (shibire^ts) to silence specific neurons (il3LN6) and measured orientation relative to the motion direction.
• Modeling: A minimal data-driven encoding model was constructed using linear response filters derived from experimental data to test if a delayed inhibition mechanism could reproduce observed direction selectivity.

Key Contributions and Results

1. Discovery of Direction-Selective Olfactory Signals: The study identified direction-selective signals in the lateral horn, the primary target of antennal lobe projection neurons. These signals exhibit a preference for "ipsi-first" motion (odor arriving at the ipsilateral antenna first) across a wide range of olfactory channels, including those processing pheromones, food odors, and aversive cues.
2. Localization of Computation: By comparing ORN terminals and ALPN dendrites, the authors demonstrated that ORN responses are largely non-directional, while ALPN dendrites in the antennal lobe exhibit robust direction selectivity. This places the computation of odor motion direction at the first synapse in the olfactory system.
3. Temporal Tuning: The direction-selective response in DA1 ALPNs is tuned to inter-antennal delays of approximately 40 ms. The system responds to naturalistic odor plume statistics, including intermittent "whiffs," and exhibits sensitivity to the sign of pairwise correlations (responding to positive correlations in the preferred direction and negative correlations in the null direction), analogous to the visual "reverse-phi" motion illusion.
4. Identification of the Circuit Mechanism: Pharmacological blockade of GABA receptors abolished direction selectivity, implicating inhibition. Connectomic analysis identified a specific GABAergic local neuron, il3LN6, as a key candidate. il3LN6 receives strong, biased input from contralateral ORNs and projects to ALPNs.
5. Physiological Validation: Physiological recordings confirmed that il3LN6 is excited by contralateral input and responds more strongly to "contra-first" motion (the null direction for downstream ALPNs). Silencing il3LN6 significantly reduced the flies' ability to orient against moving odor bars, confirming its behavioral relevance.
6. Mechanism Confirmation: A minimal data-driven model incorporating fast ipsilateral excitation and delayed contralateral inhibition (mediated by il3LN6) was sufficient to qualitatively reproduce the observed direction-selective responses in DA1 ALPNs.

Significance
This paper establishes motion detection as a circuit-level computation in the olfactory system. It reveals how the Drosophila antennal lobe transforms non-directional bilateral chemical inputs into direction-selective neural signals that guide navigation. The findings demonstrate that olfactory motion detection shares algorithmic and circuit-level logic with visual motion detection, specifically utilizing a mechanism of delayed inhibition to suppress responses to motion in the null direction. The study identifies a specific neural pathway (il3LN6 to DA1) that converts spatiotemporal chemical signals into a neural representation of odor motion, providing a foundational understanding of how animals navigate turbulent chemical environments.