Ca²⁺-Activated K⁺ Channels Reduce Network Excitability, Improving Adaptability and Energetics for Transmitting and Perceiving Sensory Information

Absence of dSK and dSlo shapes photoreceptor responses

To examine how Ca²⁺-activated K⁺ channels shape Drosophila photoreceptor voltage output, we performed in vivo intracellular recordings (Fig. 3A) from R1–R6 somata (Fig. 3B) in the retinae of dSlo⁻, dSK⁻, and dSK⁻;; dSlo⁻ null mutants and WT flies, using conventional sharp microelectrodes. Briefly, dark-adapted (∼20 s) mutant R1–R6s responded to logarithmically brightening light flashes with increasing graded depolarizations (Fig. 3C), having WT-like or slightly smaller amplitudes (Fig. 3D). However, both and dSK⁻ and dSK⁻;; dSlo⁻ R1–R6 outputs peaked faster (Fig. 3E; mean time to peak) and decayed earlier (Fig. 3F; mean half-width) to their respective resting potentials than the WT, whereas those of dSlo⁻ R1–R6s, in contrast, showed decelerated dynamics, lasting longer than the WT except at the highest intensities (Fig. 3C,F).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

R1–R6 photoreceptors of different Ca²⁺-activated K⁺ channel null mutants show distinctive response dynamics to light flashes. A, Recordings were performed in vivo from R1–R6 somata using conventional sharp microelectrodes. B, A total of 30,000 microvilli, which form a R1–R6's light-sensor, the rhabdomere (comb-like structure), sample light information (incoming photon rate changes). R1–R6 axon terminals then transmit these signals to the lamina network through sign-inverting (histaminergic) output synapses (to L1–L3 monopolar cells and amacrine cells) and receive synaptic feedback (network modulation) in return (Meinertzhagen and O'Neil, 1991; Rivera-Alba et al., 2011); the schematic highlights excitatory feedback from L2, lamina intrinsic amacrine neurons (Lai), and L4 to photoreceptor terminals. Because R1–R6s are short and have large length constants, synaptic feedback also influences their somatic response waveforms (Zheng et al., 2006; Dau et al., 2016). C, The average dSlo⁻, dSK⁻, dSK⁻;; dSlo⁻, and WT responses to 10 ms bright and dim flashes. Their corresponding delay and rise times (≤20 ms from the flash onset; light gray area) were similar, suggesting intact light information sampling. But the mutant R1–R6 responses decayed (in light brown area) either faster or slower than the WT, suggesting differences in neural tuning. D, Mutant and WT R1–R6 responses had comparable maximum amplitudes over the tested flash intensity range, resulting in similar V/log(I) saturation curves. E, dSK⁻;; dSlo⁻ and dSK⁻ responses peaked, on average, sooner than the WT to all test intensities, with significantly shorter dSK⁻ values for BG-0.5 (p = 0.028) and BG-1 (p = 0.047). Conversely, dSlo⁻ responses peaked later than the WT to all but the two brightest flashes. Moreover, these responses peaked significantly later than those of dSK⁻ at BG-0.5 (p = 0.035) and BG-3 (p = 0.013), and dSK⁻;; dSlo⁻ at BG-2 (p = 0.038) and BG-3 (p = 0.013). F, WT response half-widths to the brightest flash (BG0) were significantly longer than those of dSK⁻ (p = 0.003) and dSK⁻;; dSlo⁻ (p = 8.56 × 10⁻⁴) R1–R6s. Conversely, dSlo⁻ responses, on average, lasted the longest over a broad flash intensity range versus dSK⁻;; dSlo⁻: at BG-1 (p = 0.042), BG-2 (p = 0.011), and BG-3 (p = 0.007). D–F, Data are mean ± SEM (two-tailed t test).

Notably, however, in all the corresponding recordings, the early light-induced depolarizations (Fig. 3C, light gray area) were similar, implying that the mutant R1–R6s sampled light information normally. Thus, phototransduction reactions inside R1–R6's ∼30,000 microvilli (photon sampling units; Fig. 3B), which form its light sensor, the rhabdomere (Hardie and Juusola, 2015), seemed unaffected by the absence of Ca²⁺-activated K⁺ channels. But, instead, these mutant genotypes influenced more the subsequent neural information modulation phase (Fig. 3C, light brown area).

Response differences not from homeostatic ion channel expression

If an R1–R6 photoreceptor was an isolated system, missing Ca²⁺-activated K⁺ conductances would directly increase its membrane resistance, R_m, and consequently its time constant (τ_m = R_m · C_m; C_m is membrane capacitance). This would slow down voltage responses to light changes. However, in vivo, as each R1–R6 features complex bioelectric interactions within its membrane and with its neural neighbors, the mutant responses showed far more sophisticated dynamics (Fig. 3), presumably reflecting homeostatic changes in these interactions (Marder and Goaillard, 2006; Vähäsöyrinki et al., 2006). Therefore, to work out what made the mutant R1–R6 outputs differ, we analyzed changes both in their intrinsic (membrane) properties and extrinsic (synaptic) feedback from the surrounding network.

We first asked whether the differences in dSlo⁻, dSK⁻, and dSK⁻;; dSlo⁻ R1–R6 voltage responses resulted from homeostatic somatic conductance changes. These would affect their membrane resistances, accelerating or decelerating signal conduction. For example, missing dSK channels in dSK⁻ photoreceptors could be compensated by upregulating dSlo channel expression, for which these cells carry a normal gene; and vice versa in dSlo⁻ photoreceptors. Alternatively, the cells could increase K⁺ or Cl⁻ leak conductances (Niven et al., 2003; Vähäsöyrinki et al., 2006). While such intrinsic homeostatic mechanisms could accelerate dSK⁻ R1–R6 output, these would also lower their resting potentials: by reducing depolarizing Ca²⁺ load and/or increasing hyperpolarizing K⁺/ Cl⁻ loads. Equally, a lack of such homeostatic ion channel expression changes could have contributed to dSlo⁻ photoreceptors' slower signaling.

To test these hypotheses, we measured in vivo somatic electrical membrane properties in dark-adapted mutant and WT R1–R6s (Fig. 4A) using single-electrode current clamp (e.g., Juusola and Weckström, 1993). We found that all the mutant R1–R6s charged smaller, but broadly WT-like, voltage responses to injected current pulses (Fig. 4B). Depolarization to positive currents showed characteristic outward rectification (arrows), caused by activation of voltage-dependent K⁺ channels (Hardie, 1991a; Hardie et al., 1991; Juusola and Hardie, 2001b; Vähäsöyrinki et al., 2006), whereas hyperpolarization to negative currents, in effect, charged their membranes passively.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

In darkness, R1–R6 photoreceptors of Ca²⁺-activated K⁺ channel null mutants have lower membrane resistances and higher resting potentials. A, In vivo R1–R6 recordings. R1–R6 terminals provide histaminergic feedforward inhibition to LMCs. In return, R1–R6s receive excitatory feedback from L2 and L4 monopolar cells. B, Voltage responses of dark-adapted WT and mutant R1–R6s to intracellular current pulse injections. Arrows indicate outward rectification, caused by voltage-sensitive Shaker and Shab K⁺ conductance activation to fast membrane depolarizations. C, Mean WT R1–R6 input resistance (124.8 ± 52.7 mΩ, n = 7 cells) is significantly higher than that for all the mutant recordings (89.8 ± 34.5 mΩ, n = 47, p = 0.024), but not for each mutant type separately (dSK⁻, 74.6 ± 52.0 mΩ, p = 0.071, n = 8; dSlo⁻, 90.4 ± 33.4 mΩ, p = 0.232, n = 21; dSK⁻;; dSlo⁻, 95.9 ± 25.6 mΩ, p = 0.519, n = 18). D, Mutant R1–R6s are more depolarized than the WT photoreceptors in darkness (WT, −57.2 ± 6.1 mV; dSK⁻, −46.1 ± 3.5 mV; dSlo⁻, −47.0 ± 7.4 mV; dSK⁻;; dSlo⁻, −43.9 ± 5.8 mV). C, D, Data are mean ± SD. p values from ANOVA for multiple comparisons with Bonferroni test.

The membrane input resistances of the mutant R1–R6s (Fig. 4C), as determined by small hyperpolarizing responses to −0.02 nA current steps, were characteristically lower than in the WT (Juusola and Hardie, 2001b; Niven et al., 2003), with the mean resistance of dSK⁻ R1–R6s being the lowest (compare Abou Tayoun et al., 2011). Most crucially, however, the mutant (dSK⁻, dSlo⁻, and dSK⁻;; dSlo⁻) photoreceptors' resting potentials (Fig. 4D), instead of being more hyperpolarized, were >10 mV more depolarized than the WT. Here, if dSK⁻ or dSlo⁻ R1–R6s' intrinsic signaling properties were regulated homeostatically, by ion channel expression (as hypothesized), then their resting potential in darkness should have been below the WT range, rather than above it. Also, the higher resting potentials (Fig. 4D) and lower membrane resistances (Fig. 4C) should have accelerated signal conduction. Yet, the mean dSlo⁻ R1–R6 voltage response time-to-peak values to intermediate light flash intensities were, indeed, slower than in the WT (Fig. 3E,F).

Hence, collectively, these results suggested that the accelerated (dSK⁻ and dSK⁻;; dSlo⁻) and decelerated (dSlo⁻) light-induced voltage response dynamics of the mutant photoreceptors (Figure 3C–F) unlikely resulted from compensatory expression of leak- or Ca²⁺-activated K⁺ channels at the somata but required other/further mechanisms. However, these results were recorded from dark-adapted R1–R6s at relative rest, without LIC interactions. During brightening light stimulation, LIC and other conductances increase progressively with membrane depolarization (e.g., Song et al., 2012; Juusola et al., 2017), reducing resistance further by ≫10-fold (e.g., Juusola and Weckström, 1993), as our recordings and simulations clarify later on.

Response differences not by transduction or K⁺ conductance differences

To eliminate the possibility that developmental morphological defects in the mutant R1–R6s would have caused their altered responses, we assessed the mutant and WT eyes/retinae using both electron microscopy (Fig. 5A, top) and light microscopy (Fig. 5A, below). We found no obvious morphological differences between the eyes, with each method displaying highly ordered ommatidia with normal looking intact R1-R7 photoreceptor rhabdomeres.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

dSK⁻, dSlo⁻, and dSK⁻;; dSlo⁻ photoreceptors have normal morphology, LICs, but 8%–40% reduced light-insensitive I_A currents. A, The mutant retinae appear structurally intact, with R1–R7s having WT-like rhabdomeres and pigmentation. Scale bars (EM), 1 μm. B, Whole-cell recordings were performed from dissociated ommatidia. C, Mutant and WT R1–R6 QB waveforms and their amplitude distributions to dim flashes are similar. D, WT and dSK⁻;; dSlo⁻ R1–R6 LIC responses to 1 s light pulses of different brightness. E, Macroscopic LIC peak and plateau responses were similar, with the normal experimental variation, indicating that dSK and dSlo deletions do not affect phototransduction. The smaller dSlo⁻ maximum LIC probably resulted from these ommatidia being smaller, reflecting dSlo⁻ mutants reduced yield/health. F, WT and mutant R1–R6s' voltage-sensitive outward K⁺ currents to increased voltage steps contain both the transient Shaker (I_A) and sustained delayed rectifier, Shab (I_KS) components. G, dSK⁻ R1–R6 K⁺ currents have a reduced I_A but near normal-sized I_KS. H, On average, the maximum I_A and I_KS currents in dSK⁻ and dSlo⁻ R1–R6s are a bit smaller than the WT (dSK⁻ I_A: 36.4% < WT; I_Ks: 24.9% < WT; dSlo⁻ I_A: 30.6% < WT; I_Ks: 19.0% < WT) but WT-like in dSK⁻;; dSlo⁻ R1–R6s. n.s., non-significant.

Nevertheless, deletion of dSlo, dSK, or both could still affect intracellular [Ca²⁺] regulation, and thus potentially alter microvillar phototransduction functions indirectly (Song et al., 2012; Hardie and Juusola, 2015), modifying sampling, amplification, or integration of LICs. We therefore used whole-cell recordings in dissociated ommatidia (Hardie, 1991b) (Fig. 5B) to compare the mutant and WT R1–R6s' elementary responses (QBs) to single photons (Fig. 5C) and macroscopic LICs to light pulses (Fig. 5D,E). In this preparation, photoreceptor axon terminals were severed, cutting off any synaptic feedback from the lamina network to R1–R6s (Zheng et al., 2006).

We found the mutant R1–R6s' bump amplitudes and waveforms (Fig. 5C) and macroscopic LICs (Fig. 5D,E) to increasing light intensities WT-like, showing normal dynamics within the normal experimental variation. Here, the smaller dSlo⁻ LIC maxima likely resulted from the smaller size of these homozygotic mutant flies due to their lower yield/reduced health. Thus, deletion of dSlo, dSK, or both channels neither disrupted the microvillar R1–R6 morphology nor its phototransduction functions, again suggesting that the mutant R1–R6s would sample light information like their WT counterparts (Song et al., 2012; Hardie and Juusola, 2015; Juusola and Song, 2017).

Intriguingly, however, K⁺ conductances in dissociated dSK⁻ and dSlo⁻ R1–R6s showed slightly reduced (19%–36%) fast A (I_A or Shaker) and delayed rectifier currents (I_KS or Shab) (Fig. 5F–H), whereas these currents were broadly WT-like in dSK⁻;; dSlo⁻ R1–R6s. The decrease in the I_A and I_Ks currents together with dSK or dSlo current removal should, with other things being equal, increase membrane resistance and its time constant, leading to slower voltage responses. Instead, in vivo, we found resistance in all the mutant R1–R6s below the WT (Fig. 4C), with both dSK⁻ and dSK⁻;; dSlo⁻ R1–R6s responding faster and only dSlo⁻ R1–R6s slower (Fig. 3E), implying that homeostatic changes in K⁺ channel expression alone cannot explain their response differences.

Together, the observed normal rhabdomere morphology, WT-like LIC dynamics, and only partly reduced photo-insensitive membrane conductances implied that the mutant R1–R6s' accelerated or decelerated voltage responses, higher resting potentials, and lower membrane resistance in vivo could not be induced by homeostatic ion channel expression changes in photoreceptor somata alone. But this would more require network adaptation (Nikolaev et al., 2009; Zheng et al., 2009), parallel changes in the synaptic network activity. In such scenarios, missing one or both Ca²⁺-activated K⁺ channels would cause a homeostatic (automatic) rebalancing of the bidirectional signal transfer between photoreceptor axon terminals and the lamina interneurons (Shaw, 1984; Zheng et al., 2006, 2009; Abou Tayoun et al., 2011; Dau et al., 2016).

dSK or dSlo absence changes network adaptation

In the adult Drosophila brain, dSlo and dSK share similar expression patterns with higher expression in the lamina and medulla neuropils and weaker in the retina (Becker et al., 1995; Abou Tayoun et al., 2011). Thus, theoretically, dSlo and dSK could coparticipate in shaping the bidirectional signal transfer between R1–R6 photoreceptor axons and LMCs, which form columnar R-LMC-R network processing units in the lamina (Nikolaev et al., 2009; Zheng et al., 2009). Here, the deletion of one or the other ion channel could disrupt this balance.

We therefore next asked how Ca²⁺-activated K⁺ channels might contribute to network adaptation in the R-LMC-R system. We recorded dSK⁻, dSlo⁻, dSK⁻;; dSlo⁻, and WT R1–R6 responses to a repeated 1 s naturalistic light intensity time series stimulus (NS) (van Hateren, 1997) in vivo, and found each of them adapting differently (Fig. 6A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Adaptation dynamics and information rates of WT and Ca²⁺-activated K⁺ channel null mutant R1–R6 photoreceptors. A, Intracellular voltage responses to a repeated 1-s-long bright naturalistic light intensity time series stimulus (NS). B, Change in the response mean (± SD) over the repeated stimulation. Mean WT and dSlo⁻ R1–R6 outputs declined near exponentially to steady state, whereas adaptation in mean dSK⁻ and dSK⁻;; dSlo⁻ R1–R6 outputs depicted unique undershoots. C, Mean waveforms ± SD of steady state adapted 1 s responses. D, Relative change in R1–R6 output range, measured as the SD of the responses at each second during 50 s of stimulation (mean ± SD). WT and dSlo⁻ R1–R6s desensitized during repeated stimulation, following exponential time constants. WT R1–R6 output range contracted from 114% to 100% in ∼26 s (τ_Wild-type = 1.96 ± 0.39 s), dSlo⁻ from 134% in ∼19 s (τ_dSlo4 = 8.3 ± 1 s), and dSK⁻;; dSlo⁻ from 133% in 8 s (τ_dSK;;dSlo4 = 3.4 ± 0.4 s). dSK⁻ and dSK⁻;; dSlo⁻ R1–R6s' output ranges showed further sensitizing trends, reaching a steady state after ∼40 s. E, Mutant and WT R1–R6s' average signal-to-noise ratios, measured from their steady-state outputs to bright NS, are high and broadly similar. F, Mutant and WT R1–R6s sampled information from dim, moderately intense (middle), and bright naturalistic stimulation in a comparable manner (mean ± SD; n = 10–33 recordings). dSK⁻;; dSlo⁻ R1–R6s had a marginally lower mean information transfer rate than the other genotype photoreceptors. In each genotype, R1–R6 information rates to the given stimulation vary naturally (up to ∼200 bits/s) as each cell receives different amount of information from the network (Juusola et al., 2017). n.s., non-significant.

The mean of the WT response (Fig. 6B, black trace; measured at each second) decreased approximately exponentially as the cells adapted to NS (Fig. 6C), reaching a relative steady state in 15–20 s (Fig. 6B,C). In contrast, the corresponding means of the mutant responses declined faster but then displayed unique genotype-specific undershooting. The means of dSK⁻ (red trace) and dSK⁻;; dSlo⁻ (orange) responses first decreased to their minima in <10 s, and then increased, as the cells gradually further depolarized, reaching a relative steady state in 35–40 s; ∼20 s later than the WT, whereas the mean of dSlo⁻ photoreceptor output (blue) decayed slower than in the other mutant R1–R6s and undershot less.

Concurrently, the WT and mutant R1–R6 output ranges, measured as the SD (Fig. 6D) of their response waveforms (Fig. 6C) at each second of NS, adapted with distinctive dynamics and speeds. dSlo⁻ R1–R6 outputs desensitized the slowest, slower than the WT, with their ranges compressing with different average time courses (τdSlo⁻ = 3.41 ± 3.28 s, n = 19 cells [22 recordings]; τ_Wild-type = 1.47 ± 0.67 s, n = 7 cells [10 recordings]; mean ± SD) (Fig. 6D). Conversely, dSK⁻ and dSK⁻;; dSlo⁻ R1–R6 output ranges first compressed as rapidly as the WT (τ_dSK− = 1.45 ± 0.66 s, n = 7 cells [7 recordings]; τ_{dSK−;;Slo−} = 1.44 ± 0.32 s, n = 8 cells [9 recordings]), but then slowly began to expand, reflecting their rather similar mean voltage dynamics (Fig. 6B). The adaptive range reduction occurred most severely in dSK⁻;; dSlo⁻, and dSlo⁻ R1–R6s, leaving their steady-state responses ∼10% smaller than those of the WT.

These results highlight the complex role of Ca²⁺-activated K⁺ channels in regulating R1–R6 output in network adaptation. While the absence of dSlo channel slowed adaptation in dSlo⁻ R1–R6s, the dSK⁻ and the double mutant dSK⁻;; dSlo⁻ R1–R6s adapted faster but showed overshooting dynamics. Consequently, as an overall sign of compromised gain control, the mutant R1–R6s reached their steady-state responsiveness 20–30 s later than the WT. Thus, each mutant R-LMC-R system adapted suboptimally, constrained to its own unique dynamics.

dSK or dSlo absence leaves information sampling intact

A R1–R6's information transfer rate depends mostly on its photon-absorption rate changes, set by the number of individual sampling units (rhabdomeric microvilli) and the speed and refractoriness of their phototransduction reactions (Song et al., 2012; Juusola and Song, 2017; Juusola et al., 2017). In contrast, obeying the data processing theorem, any changes in membrane filtering affect signal and noise equally, and therefore cannot increase information (Shannon, 1948; Juusola and de Polavieja, 2003; Cover and Thomas, 2006). Accordingly, information transfer rates of mutant photoreceptors with normal phototransduction but without specific K⁺ channels, such as the slow delayed rectifier Shab (I_KS) (Vähäsöyrinki et al., 2006), are broadly WT-like. But mutations that damage ion channels can destroy information. For example, Sh mutant R1–R6s' “nonfunctional” Shaker (I_A) K⁺ channels appear to truncate signal amplification while generating noise, reducing information flow (Niven et al., 2003). Critically, however, the R-LMC-R system has intrinsic potential to combat detrimental changes within its parts. A R1–R6's impaired function can be compensated in part by extra light information (through gap-junctions and feedback synapses) from its neighbors, in which receptive fields face the same visual area (Shaw, 1984; Zheng et al., 2006; Wardill et al., 2012; Juusola et al., 2017).

Because dSlo⁻, dSK⁻, and dSK⁻;; dSlo⁻ mutant R1–R6s lack completely their functional channels (which thus should not generate extra noise) and have normal rhabdomere morphology and LIC dynamics (Fig. 5), theoretically, their somatic information transfer rates should be WT-like, or slightly lower, in case their LMC feedback was compromised.

To test this hypothesis, we compared dSlo⁻, dSK⁻, and dSK⁻;; dSlo⁻ R1–R6s' encoding performance with the WT control using the same recordings as above. In each case, the first 20–30 responses with the adapting trends were removed. The signal was taken as the average of the next 20 responses, which thus had settled to a relative steady state, with its power spectrum calculated by Fourier transform. The corresponding noise power spectrum was estimated from the difference between each response and the signal (see Materials and Methods).

We found that the mutant R1–R6s' signal-to-noise ratios (Fig. 6E) and information rates (Fig. 6F) were broadly WT-like, increasing in parallel with brightening light, as tested for dim, middle, and bright NS. Thus, as hypothesized, after the initial ∼20–30 s adaptation phase, the loss of dSK, dSlo, or both channels affected only marginally a R1–R6's encoding performance. These results highlight the R-LMC-R system's robustness and compensatory ability to withstand internal damage.

dSK or dSlo absence increases synaptic feedback

To work out in theory how synaptic feedback from the lamina interneurons should shape the WT R1–R6 output and how homeostatic feedback changes should shape mutant R1–R6 outputs, we next combined biophysical R1–R6 modeling with intracellular recordings.

Our biophysical R1–R6 model (Fig. 7A) incorporates 30,000 computational microvilli (Song et al., 2012), each of which implements full stochastic phototransduction reactions to transduce absorbed photons into QBs. Essentially, this model samples light information much like a real R1–R6 (Song et al., 2012; Song and Juusola, 2014; Juusola and Song, 2017; Juusola et al., 2017). Its QBs sum up realistic macroscopic LIC, with the best performance for naturalistic stimuli at 1–8 × 10⁵ photon absorptions/s (Song and Juusola, 2014; Juusola et al., 2017). LIC then charges a HH-type photoreceptor membrane circuit (Figs. 7B, 8) (Niven et al., 2003; Vähäsöyrinki et al., 2006; Song et al., 2012; Song and Juusola, 2014), generating output that approximates intracellular recordings to comparable light stimulation (Song et al., 2012; Song and Juusola, 2014, 2017; Juusola et al., 2017). Most differences in the simulated and recorded response waveforms would then be caused by the real R1–R6s' synaptic feedback currents, input from LMCs/amacrine neurons (Zheng et al., 2006; Rivera-Alba et al., 2011; Dau et al., 2016), which the model lacks (Juusola et al., 2017). Moreover, given that the mutant R1–R6s' phototransduction is WT-like and voltage-sensitive conductances either WT-like or only moderately reduced (Fig. 5), their voltage response differences should also mostly reflect synaptic feedback differences (Fig. 6).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Predicted synaptic feedback and ATP consumption of WT and mutant R1–R6 photoreceptors. A, Biophysically realistic R1–R6 model has four modules: stimulus generation, stochastic photon sampling/QB generation, bump integration, and voltage-sensitive membrane. But it lacks the synaptic feedback from the lamina network (Juusola et al., 2017; Song and Juusola, 2017), which affect the real R1–R6 output. R1–R6 simulations (dark yellow trace) and recordings (dotted trace) to a repeated naturalistic light intensity time series (NS) were analyzed at relative steady-state adaptation (compare Fig. 6C). B, Characteristic recording waveform to bright NS (BG0). C, Synaptic feedback to each recording was estimated computationally by linking it to the photoreceptor model, which had no free parameters. A new flat (zero) conductance, representing the synaptic input, was then injected to the model. This conductance waveform was shaped in a closed loop until the model output (gray) matched the recorded output (black). D, The fixed light-induced (dark red) K⁺ currents and the average predicted synaptic feedback and of WT and mutant R1–R6 recordings. E, Together, these currents charged up their respective simulated R1–R6 voltage responses. The simulations (light colors) match the recordings (bright colors) near perfectly. F, The average predicted synaptic feedback was unique to the mutant R1–R6s and showed stronger modulation on a higher mean (tonic excitatory background) than the WT (see also Fig. 10F). Testing the feedback means across all recordings: WT versus dSK⁻, p = 0.041; dSK⁻ versus dSlo⁻, p = 0.033; dSlo⁻ versus dSK⁻;; dSlo⁻, p = 0.009, testing the mean feedback waveforms against each other (p < 2.274 × 10⁻⁶²). G, Separately recorded LMC response waveforms to the same NS much resemble the predicted feedback waveforms in F, suggesting that L2 and L4 (and possibly amacrine neurons), which form feedback synapses with R1–R6s (Meinertzhagen and O'Neil, 1991; Rivera-Alba et al., 2011), would contribute to R1–R6 output modulation (Zheng et al., 2006). H, With these conductances included in each separate WT, dSK⁻, dSlo⁻, and dSK⁻;; dSlo⁻ R1–R6 models, the metabolic energy (ATP) consumption of each recording was calculated for its full waveform (left) (Song and Juusola, 2014) and DC voltage (right) (Laughlin et al., 1998), respectively. Notably, the original DC voltage method, which does not consider how the dynamic ion fluctuations add to the electrochemical pumping work, underestimates ATP consumption by one-third (33.2%; see Materials and Methods). I, The cost of neural information was calculated for each recording by dividing its information rate estimate with its full ATP consumption rate estimate. J, On average, the absence of dSK, dSlo, or both increased the cost of neural information in a mutant R1–R6 by 13.1% (dSK), 28.0% (dSlo), or 42.7% (dSK;; dSlo). n.s., non-significant.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

R1–R6 photoreceptors' characteristic voltage-sensitive Shaker and Shab K⁺ current responses ex vivo and their HH models to voltage commands, under whole-cell voltage-clamp conditions. A, WT. B, dSK⁻. C, dSlo⁻. D, dSK⁻;; dSlo⁻.

Therefore, we could extrapolate the synaptic feedback current to each recorded R1–R6, whether WT or mutant, computationally (Fig. 7C), by using the same fixed LIC with their specific I_A and I_SK current dynamics (Figs. 5, 8). In these simulations, we first injected a new flat (zero) conductance, representing the missing synaptic input, to the full R1–R6 model. The software then shaped up this conductance waveform in a closed loop until the model's voltage response matched the recorded response for the same light stimulus. Thus, theoretically, the resulting (predicted) current should closely mimic the real synaptic feedback, which the tested R1–R6 would have received from the lamina network in vivo.

Figure 7D shows the corresponding mean LIC and synaptic feedback estimates to repeated light stimulation for the tested WT and mutant photoreceptors, and the concurrent voltage-sensitive K⁺ currents and K⁺ leak estimates. In these simulations, while the LIC was the same (fixed; dark red traces) for every genotype, their synaptic feedback and K⁺ (dark green) currents balanced out differently to reproduce their respective in vivo voltage signals (Fig. 7E).

We found that, in every simulation, the predicted synaptic feedback to R1–R6s was excitatory, graded, and phasic (Fig. 7D,F). It rapidly increased (“switched-on”) during light decrements and decreased (“switched-off”) during light increments. This accentuated transient (phasic) light changes in photoreceptor output (Fig. 7E; compare Fig. 7A). Moreover, the predicted synaptic excitatory load to R1–R6s (Fig. 7F) was unique for each mutant and the WT flies with the highest mean to dSK⁻ (red) and dSlo⁻ (blue) photoreceptors. Thus, the enhanced excitatory feedback conductance from the lamina interneurons is the most probable mechanistic explanation of why and how the mutant photoreceptors were more depolarized than their WT counterparts, both in darkness (compare Fig. 4D) and during light stimulation (Fig. 7E).

Remarkably, these feedback dynamics (Fig. 7F), which were extrapolated using only photoreceptor data (Fig. 7A–C), closely resembled postsynaptic intracellular LMC responses to the same light stimulus (Fig. 7G). This implied that L2, L4, and lamina intrinsic amacrine neurons (Lai), all of which receive inhibitory inputs from R1–R6 but form excitatory feedback synapses to R1–R6 (Kolodziejczyk et al., 2008; Raghu and Borst, 2011; Hu et al., 2015; Davis et al., 2018), could alone or together be the major source of this feedback. Thus, these new findings are consistent with our theory of how the R-LMC-R system, by dynamically balancing its inhibitory and excitatory synaptic loads, shapes the early neural representation of visual information (Zheng et al., 2006, 2009; Nikolaev et al., 2009; Dau et al., 2016).

dSK and dSlo lower neural information energy cost

In response to LIC and synaptic feedback, ion channels open and close, regulating the ionic flow across the photoreceptor membrane. Meanwhile, its ion cotransporters, exchangers, and pumps uptake or expel ions to maintain ionic concentrations inside and outside. The work of the pumps in moving ions against their electrochemical gradients consumes ATP (Laughlin et al., 1998). For R1–R6, a reasonable estimate of this consumption can be calculated from the ionic flow dynamics through its ion channels (see Materials and Methods) (see also Song and Juusola, 2014).

Using our biophysical R1–R6 model, which now included the synaptic feedback, we calculated how much each recorded WT and mutant R1–R6 consumed metabolic energy (ATP molecules/s) to encode bright naturalistic light changes (Fig. 7H, left). We discovered that, because their enhanced synaptic feedback held dSK⁻, dSlo⁻, and dSK⁻;; dSlo⁻ R1–R6s at higher operating voltages, where signaling is more expensive, they consumed on average 13.3%, 18.3%, and 10.2% more ATP than the WT, respectively.

We also estimated each tested R1–R6's ATP consumption by using the method of balancing out the ionic currents for its light-induced mean (flat) depolarization level, or DC (Laughlin et al., 1998). This produced a metric, which followed quite a similar trend (Fig. 7H, right). But because it discarded how much the dynamic ion fluctuations increase the work to maintain transmembrane ionic concentration, it underestimated the total ATP consumption by approximately one-third.

Next, using the full biophysical models (Fig. 9), we calculated how the mutant R1–R6s' homeostatically reduced Shaker and Shab K⁺ conductances (Fig. 5F–H) affect their neural information costs. We fixed the Shaker and Shab conductance dynamics of the dSK⁻, dSlo⁻, and dSK⁻;; dSlo⁻ R1–R6 models to match typical WT R1–R6 VC recordings (Fig. 8A). This increased the mutant photoreceptors' energy consumption, but only slightly (compare Fig. 9H with Fig. 7H). Hence, the observed homeostatic 19%–36% Shaker and Shab current reduction in dSK⁻ and dSlo⁻ R1–R6s (Fig. 5F,H) made evolutionary sense, as it cut both their hyperpolarizing drive, which therefore would require less excitatory synaptic feedback to depolarize the cells, and neural information costs. But this saving was small, only 4.5%–6.2%. And somewhat unexpectedly, its homeostatic effect, indeed, increased the dSK⁻ and dSlo⁻ R1–R6s' synaptic feedback overload slightly with respect to dSK⁻;; dSlo⁻ R1–R6s, which had WT-like Shaker and Shab conductance dynamics (Fig. 7F).

• Download figure
• Open in new tab
• Download powerpoint

Figure 9.

Examples of the WT (black), dSK⁻ (red), dSlo⁻ (blue), and dSK⁻;; dSlo⁻ (orange data) photoreceptor models' Shaker (I_A), Shab (I_Ks), K⁺ leak, LIC, and synaptic feedback currents to naturalistic light stimulation, when using the same WT Shaker and Shab conductance dynamics in all the models (compare Figs. 7, 10). A, In these simulations, the WT, dSK⁻, dSlo⁻, and dSK⁻;; dSlo⁻ R1–R6 models had identical LIC and voltage-sensitive membrane conductances (the models used the WT Shaker and Shab dynamics as in Fig. 8A; compare Fig. 7A). B, Characteristic recording waveform to bright NS (BG0). C, Again, synaptic feedback was computed through the R1–R6 model, which had no free parameters, in a closed loop until the model output (gray) matched the recorded output (black). D, The fixed light-induced (dark red) K⁺ currents and the average predicted synaptic feedback of WT and mutant R1–R6 recordings. E, Together, these currents charged up their respective simulated R1–R6 voltage responses. The simulations (light colors) matched the recordings (bright colors). F, Again, the average predicted synaptic feedback to the mutant R1–R6s were stronger, having higher means (tonic excitatory background) than the WT. G, The recorded LMC response waveforms to the same NS resembled the predicted feedback waveforms in F. H, The ATP consumption of these mutant R1–R6 models was 3.5%–8.6% higher than in those models, which had their recorded (smaller) Shaker and Shab conductances (compare Fig. 5F–H and Fig. 7H). I, The cost of neural information was calculated for each recording by dividing its information rate estimate with its full ATP consumption rate estimate. J, On average, the absence of dSK, dSlo, or both increased the cost of information in a mutant R1–R6 by 19.333.4% (dSK), 32.628.6% (dSlo), or 43.919.3% (dSK;; dSlo). Thus, 19%–36% homeostatic reductions in Shaker and Shab currents in dSK and dSlo R1–R6 photoreceptors caused only 4.5% and 6.2% savings in their ATP consumption per each bit of transmitted information (compare Fig. 7I,J). n.s., non-significant.

Furthermore, simulations about other possible homeostatic changes (Fig. 10) indicated that, by increasing leak and voltage-sensitive K⁺ conductances, or adding an extra Cl⁻ leak, in the R1–R6 membrane would strengthen and accentuate synaptic feedback (Fig. 10F), and by that increase both the WT R1–R6s' ATP consumption (Fig. 10H; now by 23.2%) and the mutant photoreceptors' neural information costs in respect to the WT (Fig. 10I,J), now by 22.3% (dSK⁻), 37.0% (dSlo⁻), or 57.6% (dSK⁻;; dSlo⁻). Therefore, as energy wasting reduces fitness, the earlier proposed leak conductance overexpression alone (Niven et al., 2003; Vähäsöyrinki et al., 2006) seems an unlikely homeostatic strategy here.

• Download figure
• Open in new tab
• Download powerpoint

Figure 10.

Examples of the WT (black), dSK⁻ (red), dSlo⁻ (blue), and dSK⁻;; dSlo⁻ (orange data) photoreceptor models' Shaker, Shab, K⁺ leak, Cl⁻ leak, LIC, and synaptic feedback currents to naturalistic light stimulation, when using larger WT Shaker and Shab conductance dynamics (as in Table 2) in all the models (compare Figs. 7 and 9). A, In these simulations, we further added a Cl⁻ leak and Cl⁻ conductance (here combined to g_Cl− to simply the HH diagram) in the R1–R6 photoreceptor membrane model and balanced these by increasing voltage-sensitive K⁺ conductances (see Materials and Methods; speculative photoreceptor membrane model), and again the synaptic feedback (Juusola et al., 2017; Song and Juusola, 2017) was computed in a closed loop until the simulations matched the recordings (compare Fig. 6C). B, Characteristic recording waveform to bright NS (BG0). C, Synaptic feedback to each recording was estimated computationally by linking it to the photoreceptor model, which had no free parameters. D, The fixed light-induced (dark red) K⁺ and Cl⁻ currents and the average predicted synaptic feedback and of WT and mutant R1–R6 recordings. E, These currents charged up their respective simulated R1–R6 voltage responses (light colors), which matched the actual recordings (bright colors). F, Similar to the other simulations (compare Figs. 7E and 9E), the predicted synaptic feedback to the mutant R1–R6s was larger (carrying bigger modulation) with a higher mean (tonic excitatory background) than the WT. However, the modulation in these simulations was even more transient. G, Separately recorded LMC response waveforms to the same NS much resemble the predicted feedback waveforms in F. H, Energy (ATP) consumption of each recording was calculated for its full waveform (left) (Song and Juusola, 2014) and DC voltage (right) (Laughlin et al., 1998), respectively. Notably, the added extra Cl⁻ leak and Cl⁻ conductance (with rebalanced K⁺ conductances) increased the photoreceptors' energy usage by ∼26.3% (compare Fig. 7H, left: from 2.10 × 10⁹ ATP/s to 2.65 × 10⁹ ATP/s), whereas the original DC voltage method (right), which does not consider how the dynamic ion fluctuations add to the electrochemical pumping work, now underestimated ATP consumption by ∼15%. I, The cost of neural information was calculated for each recording by dividing its information rate estimate with its full ATP consumption rate estimate. J, Here, homeostatic increase in K⁺ and Cl⁻ leak conductances (to compensate the loss of Ca²⁺-activated K⁺ channels) increased the cost of information in a mutant R1–R6 by 22.2 ± 42.3% (dSK⁻), 37.0 ± 30.3% (dSlo⁻), or 57.6 ± 28.8% (dSK⁻;; dSlo⁻), in respect to the comparable WT model. n.s., non-significant.

These results establish the extra energy, which a mutant R1–R6 must spend to function without Ca²⁺-activated K⁺ channels, as a major cost for homeostatic compensation of neural information (Fig. 7I). To maintain similar information rates (Fig. 6F), an average mutant R1–R6 consumed at least 13.1% (dSK⁻; p = 0.114), 28.0% (dSlo⁻; p = 0.016), or 42.7% (dSK⁻;; dSlo⁻; p = 9.56 × 10⁻⁴) more ATP for each transmitted bit than its WT counterpart (Fig. 7J). Notably, these costs would only increase further if homeostatic compensation of the missing dSK and dSlo channels further entailed overexpression of additional K⁺ or Cl⁻ conductances or leaks (Figs. 9, 10). Thus, in Drosophila photoreceptors, Ca²⁺-activated K⁺ channels reduce the energy cost of neural information.

dSlo and dSK coregulate feedforward transmission to LMCs

Thus far, we have provided experimental and theoretical evidence that both BK (dSlo) or dSK channel deletions enhance synaptic feedback from the lamina interneurons to R1–R6s (Figs. 3–10). But these results still leave open the corresponding changes in the postsynaptic LMC output, which initiates the motion vision pathways to the fly brain (Joesch et al., 2010; Wardill et al., 2012). To test how dSK and dSlo deletions affect such feedforward transmission directly, we recorded intracellular voltage responses of dark-adapted LMCs in the mutant and WT laminae to brightening light flashes, which covered a 4-log intensity range (Fig. 11A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 11.

LMC output in the WT and mutant flies differ consistently. A, Top, Intracellular LMC recordings were performed from in vivo. Bottom, Schematic represents synaptic feedback from L2/AC and L4 to photoreceptor terminals. B, Voltage responses of the WT and mutant LMCs to bright and middle intensity flashes. C, dSlo⁻ and dSK⁻;; dSlo⁻ LMCs generated smaller responses to the dim and middle intensity flashes than the WT (dSlo⁻: p < 0.01, BG0–4, n = 8; dSK⁻;; dSlo⁻: p < 0.03, BG0–4, n = 5). D, dSK⁻ LMC responses peaked the fastest, whereas dSlo⁻ and dSK⁻;; dSlo⁻ LMC responses were slower than the WT (p < 0.03) over the tested intensity range. E, dSK⁻ LMC responses lasted as long as the WT responses, whereas dSlo⁻ and dSK⁻;; dSlo⁻ LMCs took much longer to repolarize than WT (dSlo⁻, p < 0.04, BG2–3; dSK⁻;; dSlo⁻, p < 0.04, BG0–3). F, In vivo ERGs, depicting global light-induced eye activity, were recorded from the corneal surfaces of intact Drosophila. G, dSK⁻, dSlo⁻, and dSK⁻;; dSlo⁻ ERGs often showed characteristic oscillations after the light on- and off-transients, consistent with the corresponding intracellularly recorded light-induced LMC oscillations (B). H, Dynamic spectra of the ERGs (G) reveal the frequency dependency and duration of the oscillations. C–E, Data are mean ± SEM (two-tailed t test).

Expectedly, light rapidly hyperpolarized LMCs and darkness depolarized them (Fig. 11B) (Zettler and Järvilehto, 1973; Juusola et al., 1995a; Zheng et al., 2006), driven by the photoreceptors' inhibitory transmitter, histamine (Hardie, 1989; Dau et al., 2016). Yet, these dynamics varied somewhat systematically between the genotypes, with the mutant LMCs often showing oscillating responses (ringing) around specific frequencies. L1 (on-pathway) and L2 (off-pathway) responses are thought to be largely similar at the dendritic (lamina) level (Hardie and Weckström, 1990; Uusitalo et al., 1995a; Nikolaev et al., 2009) (compare Fig. 7G), with their medulla terminals' light-on and -off preference (Joesch et al., 2010; Freifeld et al., 2013) most likely arising through specific medulla circuit processes. Therefore, with most penetrations likely from L1 and L2, which are the largest LMCs, our recordings should mostly depict mutation-induced variations and less LMC-type-dependent differences.

dSK⁻ LMC output was consistently the most transient, even to dim flashes (Fig. 11B–E), showing accelerated (most “light-adapted”) dynamics with the fastest time-to-peak values (Fig. 11D). By and large, the size (Fig. 11C) and half-width (Fig. 11E) of these responses were WT-like, but, unlike the WT, they often showed rapid oscillation bursts to dim flashes (see also Abou Tayoun et al., 2011).

In contrast, both dSlo⁻ and dSK⁻;; dSlo⁻ LMC responses to dimmer flash intensities were on average smaller than those of the WT and dSK⁻ LMCs (Fig. 11B,C). But as their amplitudes increased with light intensity, the brightest flashes evoked about the same size responses from all the genotypes (Fig. 11B,C). Therefore, during dim (but not bright) stimulation, the excitatory feedback from L2 and L4 cells to R1–R6s (Zheng et al., 2006), if directly following the recorded dSlo⁻ and dSK⁻;; dSlo⁻ LMC responses, could be driven by smaller dynamic modulation (Fig. 11C) on a larger static load (as their mean LMC responses would thus also be more depolarized). This would reduce R1–R6 membrane impedance and, presumably, synaptic gain in R1–R6 output, consistent with the smaller dSlo⁻ and dSK⁻;; dSlo⁻ R1–R6 responses to dim naturalistic light stimulation (Fig. 6C). Furthermore, dSK⁻;; dSlo⁻ LMC response dynamics were also slower and less tightly time-locked (Fig. 11D), often ringing sluggishly (Fig. 11B), prolonging the response half-width (Fig. 11E), and peaking later than the other corresponding LMC responses (Fig. 11D). Such desynchrony would add noise in the synaptic feedback, and may have contributed to the slightly lower signal-to-noise ratios and information transfer rates of dSK⁻;; dSlo⁻ R1–R6s (Fig. 6E).

Thus, deletion of dSK, dSlo, or both led to suboptimal network adaptation in the R-LMC system, seen as accelerated or decelerated LMC responses and mutation-specific oscillations. Crucially, these oscillations, with their characteristic frequencies, were also regularly observed in the mutant eyes' global electrical activity (ERGs) (Fig. 11F–H), supporting the intracellular results. Nevertheless, the observed differences cannot be directly attributed for missing dSK, dSlo, or both in the mutant LMCs (compare Abou Tayoun et al., 2011). The respective functional channels (dSK in dSlo⁻ mutants, dSlo in dSK⁻ mutants, and both channels in WT) could act remotely in the circuit, or their LMC response dynamics could result from combinatorial effects on both R1–R6s and LMCs.

Mutants' optomotor responses reflect early vision defects

To test whether the mutation-specific network adaptations influence visual perception, we measured the flies' optomotor behavior in a classic flight simulator system (Fig. 12A). Notably here, dSK⁻, dSlo⁻, and dSK⁻;; dSlo⁻ mutants lack their respective channel activity throughout their brains, and thus are likely to have perceptual deficits beyond their distorted LMC inputs, and in case of dSlo⁻, reduced health/motility (see Materials and Methods). But it is the LMC input, which sets their absolute motion vision limit (Rister et al., 2007; Joesch et al., 2010; Wardill et al., 2012). So, while any observed phenotype is convolving the channel contributions in some complex, unknown way across cell types, LMC input to the mutant brain is still its motion vision bottleneck, driving optomotor behaviors in a closed loop. Therefore, as a mutant's optomotor behavior cannot be better than, and must ultimately reflect, its LMC input, it is informative to compare their respective optomotor response with their LMC input at different stimulus conditions to determine the generic behavioral differences across the different phenotypes. Moreover, these comparisons tell us further each mutant phenotype's capacity to compensate its specific mutation effects.

The tethered WT and mutant flies generated yaw torque by attempting to follow left and right rotating panoramic scenes, which showed either coarse (14.4°) or fine-grained (3.9°) vertical black-and-white stripe patterns, facing the flies. The resulting optomotor response waveforms and sizes were used to quantify how well individual flies and their respective populations (genotypes) saw these scenes rotating either slowly (45°/s) or fast (300°/s). Although the average interommatidial angle (the eyes' optical limit) is 4.5° (Gonzalez-Bellido et al., 2011), photomechanical photoreceptor microsaccades enable Drosophila to see much finer (hyperacute) details (Juusola et al., 2017).

We found that flies of each genotype could follow these stimuli (Fig. 12A), indicating that their visual systems represented, and motor systems reacted to, the opposing (left and right) image motion appropriately. However, the relative optomotor response sizes (Fig. 12B) and waveforms (Fig. 12C) showed genotype-specific sensitivities, or stimulus preferences, which were both repeatable and independent of the stimulus presentation order. Thus, these response differences could not be caused by stimulus salience, neural habituation, or flight muscle fatigue.

• Download figure
• Open in new tab
• Download powerpoint

Figure 12.

WT flies and Ca²⁺-activated K⁺ channel mutants differ in their sensitivity to track different field rotations. A, Optomotor responses to slow (45°/s) and fast (300°/s) left/right rotating fields of either coarse (14.4°) or fine-grained (hyperacute: 3.9°) black-and-white vertical stripes (100% contrast), as seen by the flies. Top, Mean (thick line) and 19–32 responses (thin lines) of the same fly to the four stimuli. Bottom, Population means of many flies of the same genotype. Each fly was tested with these four stimuli to work out its relative optomotor sensitivity. In the experiments, the stimulus presentation order was actively varied to reduce novelty, flight fatigue or adaptation-induced bias. B, The relative output (%) range (maximum − minimum) of each tested fly, as scaled by its maximum optomotor response to the four stimuli. WT, dSK⁻, and dSK⁻;; dSlo⁻ flies typically responded the strongest to the fast coarse field rotation, whereas dSlo⁻ flies were most sensitive to slow hyperacute field rotations. C, Normalized responses of each genotype to each test stimuli. D, The relative output range (maximum − minimum) of the different genotype of each test stimuli. E, F, Normalized responses of the different genotypes, compared with each other, for each test stimuli. * highlights fast body saccades. n.s., non-significant.

WT flies preferred, on average, the fast coarse stripe field rotations (Fig. 12B, black; 96.6 ± 8.5% maximum response, mean ± SD, n = 15 flies) over the slow coarse (87.8 ± 12.6%) and slow hyperacute (66.1 ± 15.2%) stimuli, but only just. Even their responses to fast hyperacute rotations were substantial (28.9 ± 9.0%), consistent with Drosophila's high visual acuity, even at saccadic speeds (>200°/s) (Juusola et al., 2017). Such an all-round optomotor performance over a broad motion stimulus range is consistent with the idea that the optomotor behavior scales with the sensory input strength and dynamics from the eyes (Wardill et al., 2012). Thus, the adaptive signal scaling in the early visual system, seen as amplitude and time-normalized LMC contrast responses to different stimulus intensity and speed conditions (Zheng et al., 2006, 2009), would be a prerequisite for consistent perception (optomotor behavior) over a broad image motion range.

In contrast, dSK⁻ mutants responded far more strongly to the fast coarse rotating field (Fig. 12B, red; 99.8 ± 7.6% maximum response, n = 13 flies) than the other stimuli (19.9%–68.0%), with their slow and fast hyperacute field rotation responses being significantly weaker than those of the other genotypes (Fig. 12E). Interestingly and distinctively, the dSK⁻ responses were further dominated by large and fast body saccades (Fig. 12A, asterisk), which appeared at seemingly regular intervals from the stimulus onset onwards and could make >50% of their total amplitude (Fig. 12D). Thus, the accelerated dSK⁻ photoreceptor and LMC dynamics (compare Figs. 3C, 11C), and tendency to oscillate seem preserved in the dSK⁻ visual system, with these motion perception distortions possibly compelling their “spiky” optomotor responses.

The optomotor behavior of dSlo⁻ mutants showed similarly suggestive correlations to their R-LMC-R network adaptation dynamics. These flies, which boast slightly decelerated photoreceptor (Fig. 3E) and LMC (Fig. 11D) dynamics, preferred slow field rotations and, surprisingly, were most sensitive to the slow hyperacute stimulus (Fig. 12B, blue; 94.8 ± 9.0%, n = 3 flies). Although dSlo⁻ mutants, in absolute terms, generated the weakest flight simulator torque responses of the tested genotypes, the mutants that flew did so over the whole experiments, making these stimulus preferences genuine.

Finally, the sensitivity of dSK⁻;; dSlo⁻ mutant responses (Fig. 12B, orange) followed the average of dSK⁻ and dSlo⁻ mutants' optomotor responses (Fig. 12B, purple dotted line) more closely than the mean WT responses (black). In particular, their responses were relatively more sensitive to hyperacute stimuli than the corresponding WT responses (Fig. 12E) but rose and decayed slower (Fig. 12F, arrows), consistent with dSK⁻;;dSlo having slower LMC dynamics (Fig. 11D,E). Thus, suggestively, their optomotor dynamics differences reflected more differences in early visual network adaptations rather than in other systems, such as the sensorimotor.