Table 2.
Parameter | Value | Units | Role | Source |
---|---|---|---|---|
phase_shiftleft_PFL3 | −90 | deg | Anatomical shift of left PFL3 neurons entering FB | 31 |
phase_shiftright_PFL3 | 90 | deg | Anatomical shift of right PFL3 neurons entering FB | 31 |
phase_shiftPFL2 | 180 | deg | Anatomical shift of PFL2 neurons entering FB | 31 |
phase_shiftleft_PFN | 45 | deg | Anatomical shift of left PFN neurons entering FB | 31 |
phase_shiftright_PFN | −45 | deg | Anatomical shift of right PFN neurons entering FB | 31 |
peakleft | −45 | deg | Peak wind direction tuning of left PFNs | 21 |
peakright | 45 | deg | Peak wind direction tuning of right PFNs | 21 |
optogenetic_power | Low: 0.75, High: 2.3 | Activity | Activation of hΔC neurons by optogenetic stimulus | Optimized to match behavior |
khΔC | 0.1 | – | Slope of hΔC activation function | 72 |
θhΔC | 0.7 | Input | Threshold of hΔC activation function | Optimized to match behavior |
taun | 10 | ms | Time constant for exponentially filtered white noise | 72 |
σn | 0.03 | Activity | Standard deviation of exponentially filtered white noise | 72 |
Wii | I <-> i+4: 1.1 Otherwise: 0 | – | Synaptic strength between mutually inhibited neurons | 72 |
g | 0.5 | – | Adaptation weight for mutual inhibition neurons | 72 |
tauu | 1 | ms | Time constant for mutual inhibition neurons | 72 |
taua | 100 | ms | Time constant for slow feedback adaptation | 72 |
kmi | 0.1 | – | Slope of mutual inhibition activation function | 72 |
θmi | −0.15 | Input | Threshold of mutual inhibition activation function | Adjusted from72 |
taupfl | 1 | ms | Time constant for PFL3 and PFL2 neurons | 72 |
kpfl | 0.1 | – | Slope of PFL activation function | 72 |
θpfl | 1.7 | Input | Threshold of PFL activation function | Optimized to match behavior |
λ | 0.0003 | Rate | Baseline turn rate for 20000 Hz simulated data | 10 |
σturn | 20 | deg/s | Standard deviation of angular velocity distribution | 10 |
m1 | 0.03 | deg/(s*activity) | Constant coupling PFL3 activity to angular velocity at 20,000 Hz | Optimized to match behavior |
vbase | 6 | mm/s | Baseline groundspeed | 10 |
m2 | 0.25 | mm/(s*activity) | Constant coupling PFL2 activity to groundspeed | Optimized to match behavior |
Note: many parameters were manually fit to match simulation data with behavioral trajectories. While not explored in this study, many of these parameters (e.g., activation thresholds) are flexible and can be adjusted together to keep the steering circuit functioning across a range of sinusoid amplitudes.