Oscillatory brain activity in spontaneous and induced sleep stages in flies

Abstract

Sleep is a dynamic process comprising multiple stages, each associated with distinct electrophysiological properties and potentially serving different functions. While these phenomena are well described in vertebrates, it is unclear if invertebrates have distinct sleep stages. We perform local field potential (LFP) recordings on flies spontaneously sleeping, and compare their brain activity to flies induced to sleep using either genetic activation of sleep-promoting circuitry or the GABA_A agonist Gaboxadol. We find a transitional sleep stage associated with a 7-10 Hz oscillation in the central brain during spontaneous sleep. Oscillatory activity is also evident when we acutely activate sleep-promoting neurons in the dorsal fan-shaped body (dFB) of Drosophila. In contrast, sleep following Gaboxadol exposure is characterized by low-amplitude LFPs, during which dFB-induced effects are suppressed. Sleep in flies thus appears to involve at least two distinct stages: increased oscillatory activity, particularly during sleep induction, followed by desynchronized or decreased brain activity.

Fig. 1

Increased 7–10 Hz oscillations during spontaneous sleep. a In vivo overnight LFP recording setup (see Methods section). b Behavioral responses to a mechanical stimulus, in relation to prior immobility time. Top: three sample traces. Colored bars on the x-axis indicate the time period bins used for calculating response proportion. Bottom: Average response (± s.e.m.) for four prior immobility durations (n = 7, *p < 0.05, **p < 0.01 by Friedman test with Dunn’s multiple comparisons between all immobility durations). c Left: spectrogram of LFP amplitude (0–40 Hz power, see Methods section) of a sample fly recording over 24 h (top), with corresponding raw LFP signal (middle) and behavioral activity quantified as pixel changes (bottom). Right panels show expanded views of a 5-min segment of a sleep epoch (black box) and a 5-min segment of a wake epoch (orange box). White arrows indicate some instances of ~8 Hz oscillations. d Average 0–100 Hz LFP power (± s.e.m.) during nighttime and daytime sleep is significantly reduced compared to daytime wake (n = 10, *p < 0.05, ***p < 0.001 by Friedman test with Dunn’s multiple comparisons between all conditions). e Average 6–10 Hz power spectra for sleep and wake states during day and night (n = 10 flies, same color code as in d). f Sleep bouts (> 5 min) were binned into 5 segments (1 min each, except for mid-sleep, which was variable in length) to compare LFPs from early to late sleep. g Average 7–10 Hz power (± s.e.m.) for each sleep epoch, normalized to mid-sleep. (n = 10 flies, *p < 0.05, **p < 0.01 by Friedman test with Dunn’s multiple comparisons between each sleep segment and mid-sleep). Images: Melvyn Yap

Fig. 2

Thermogenetic sleep induction increases 2–40 Hz power in the central brain. a Left image shows the expression pattern of 104y-Gal4 circuit as visualized by green fluorescent protein (GFP) expression (green). The synaptic marker nc82 highlights neuropil structures (magenta). Middle and right panels indicate the approximate locations of the 16 channel probe (15 functional channels) in a standardized Drosophila brain. The channels were grouped into three regions: two optic lobes region (left and right, OL), and the central brain (CB). Scale bar = 100 μm. b Spectrograms show 2–40 Hz LFP power across all 15 channels (grouped as OL and CB) in 3-sec segments associated with heat conditions and then concatenated, for an individual fly expressing 104y-Gal4/UAS-TrpA1 (Trp/104y, top), the combined median power for Trp/104y (n = 7, middle), and 104y-Gal4/+ (+/104y, n = 7, bottom). Blue and red bars at the bottom indicate the temperature stimulus. c The 15 recording channels were grouped into the 3 aforementioned regions for the purpose of analysis. Bar plots shows the median 2–40 Hz LFP power during the periods of Heat ON (red) and Post Heat (blue) relative to values at baseline, which was normalized to zero. Comparisons were made for an individual UAS-TrpA1/104y-Gal4 fly (top), the combined averaged power for UAS-TrpA1/104y-Gal4 (middle, n = 7, *p < 0.0125), and 104y-Gal4/+ controls (bottom, n = 7, ns). UAS-TrpA1/+ controls were also tested (not shown) and no significant effects of heating were found (n = 7). Statistical significance was determined by multi-factor ANOVA with post hoc contrasts on a three-way interaction term between brain regions, fly line, and heat condition (Supplementary Note 1 and Supplementary Tables 1 and 2). Sample sizes indicate the number of flies tested. Images: Angelique Paulk

Fig. 3

Oscillations of 7–10 Hz induced in dFB. a In vivo exposed-brain LFP recording setup optimized for thermogenetics (see Methods section). b Close-up showing the posterior head (top), and with part of cuticle removed exposing the brain (bottom). Scale = 0.5 mm. c GFP expression of 23E10-Gal4 (green) highlighting the dFB. nc82 (magenta) highlights neuropil. Arrow, LFP recording site. Scale = 50 μm. d Spectrogram of 2–15 Hz LFP power (top) for a 12-min LFP recording in the dFB in 104y-Gal4,UAS-mCD8::GFP/UAS-TrpA1 (TRPA1/104y), with corresponding behavioral activity (Δ pixel, middle) and brain perfusion temperature (bottom). Dark vertical bars are excluded artifact. e Spectrograms of averaged 2–15 Hz LFP power for dFB recording in TRPA1/104 y flies (n = 8, top) and control strain 104y-Gal4,UAS-mCD8::GFP/+(+/104 y, n = 6, bottom). Bottom bar indicates time when bath temperature exceeded 29 °C (Heat ON, red). f Average percentage (± s.e.m.) of time fly spent moving during baseline, Heat ON, and Post Heat conditions for TRPA1/104y flies (top, n = 22, by Friedman test with Dunn’s multiple comparisons), and for +/104y flies (bottom, n = 17, *p < 0.05, Friedman test with Dunn’s multiple comparisons). g Median 7–10 Hz LFP power in the dFB during Heat ON for TRPA1/104y flies (top, n = 8, *p < 0.05, one sample t test comparing to baseline of zero for Heat ON), and for +/104y flies (bottom, n = 6, ns, by one-sample t test comparing to baseline of zero for both Heat ON and Post Heat). h Spectrogram of 2–15 Hz LFP power for recording in the optic lobe (OL) of a sample TRPA1/104y fly. i Spectrogram of average 2–15 Hz LFP power for OL recording in TRPA1/104y flies (n = 10). j Left, median 7–10 Hz LFP power in the OL during Heat ON for TRPA1/104y flies (left, n = 10, *p < 0.05 by one sample t test comparing to baseline of zero for Heat ON condition, and ns by Wilcoxon signed rank test comparing to baseline of zero for Post Heat condition); right, median 7–10 Hz LFP power for +/104y flies (right, n = 10, ns by one sample t test comparing to baseline of zero for Heat ON and Post Heat). ns, not significant.  Images: Melvyn Yap

Fig. 4

Gaboxadol-induced sleep is associated with an overall decrease in LFP activity. a In vivo exposed-brain LFP recording setup optimized for pharmacological experiments (see Methods section). b Spectrogram of 2–40 Hz LFP power (top) taken from the dFB of a fly exposed to Gaboxadol (0.1 mg/ml) for 5 min (magenta shade), with corresponding filtered LFP signal (middle) and corresponding behavioral activity quantified as pixel changes (bottom). c Median percentage of time fly spent moving within 5 min prior (blue) and 5 min after (red) the pre-determined drug onset (see Methods section) (*p < 0.05, by Wilcoxon matched pairs signed rank test between pre- and post-drug). d Representative fly movement (pixel change) for each concentration of Gaboxadol displaying latency to sleep for each Gaboxadol concentration (indicated on the left). e Averaged latency to sleep ( ± s.e.m.) was significantly earlier for flies exposed to a Gaboxadol concentration of 0.2 mg/ml (n = 6) compared to both 0.1 mg/ml (n = 6, **p < 0.01) and 0.05 mg/ml (n = 6, *p < 0.05 by Kruskal–Wallis with Dunn’s multiple comparisons between all concentrations). f A significant decrease in the overall LFP power (0–100 Hz) was observed when flies were exposed to Gaboxadol 0.2 mg/ml (n = 6) but not for lower concentrations (*p < 0.05 by Wilcoxon matched pairs signed rank test between pre- and post-drug). g Spectrograms of individual fly LFP recordings starting 1 min after the onset of Gaboxadol perfusion, for three concentrations of Gaboxadol. White arrows indicate 7–10 Hz oscillations. Sample sizes indicate the number of flies tested. Images: Melvyn Yap

Fig. 5

Low-frequency oscillations generated from activating the dorsal fan-shaped body (dFB) are abolished by Gaboxadol. a Target neurons for optogenetic activation (top), using the dFB-specific driver 23E10-Gal4 as visualized by green fluorescent protein (GFP) expression (green). Scale bar = 100 μm. In vivo exposed-brain LFP recording setup as in Fig. 3a, optimized for optogenetic experiments (bottom, see Methods section). b Spectrogram of a fly expressing 23E10-Gal4/UAS-CsChrimson (Chrimson/23E10) showing the presence of 7–10 Hz oscillation when dFB was optogenetically activated (top) and the averaged spectrogram (n = 7, bottom). Blue vertical bars represent excluded data due to the presence of external artifact. Bottom bar indicates when the light stimulus was on for all experiments (LED ON, red). c Photostimulation of CsChrimson-expressing dFB neurons was associated with a significant increase in the averaged 6–15 Hz and 15–30 Hz LFP power in the dFB (orange, n = 7, *p < 0.05 by Wilcoxon signed rank test comparing to baseline of zero), while this effect was not observed in the control strain 23E10-Gal4/+ (+/23E10, black, n = 4, ns by Wilcoxon signed rank test comparing to baseline of zero). d Top: experimental timeline indicating the time point of delivery of optogenetic stimulation, once prior to the delivery of Gaboxadol (0.2 mg/ml), and repeated once thereafter. Prior to exposure to drug, there was a significant increase in the average 2–6 Hz LFP power when optogenetically stimulated (bottom left, n = 8, *p < 0.05, by Wilcoxon matched pairs signed rank test between pre and post drug). No significant changes to the average LFP power of any frequency domain were detected with optogenetic stimulation after the drug was delivered (bottom right, n = 8, ns, by Wilcoxon matched pairs signed rank test between pre and post drug). Sample sizes indicate the number of flies tested. Images: Melvyn Yap

Fig. 6

Behavioral effects and consequences of dFB sleep vs. Gaboxadol sleep. a Flies in glass tubes were filmed from above for the duration of the experiment. DART software was used to track fly activity and test behavioral responsiveness using a mechanical vibration. Red LEDs were used for optogenetic activation (see Methods section). b Behavioral responsiveness either probing for general responsiveness or arousal thresholds (top right panels) was tested by quantifying the change in fly locomotor activity following the vibration stimulus (measured in g, see Methods section). Following stimulus delivery (dashed red line), flies increase their locomotion speed as shown by their displacement in the tube (top left panel). Responsiveness could be binned by prior immobility groups, where > 5 min of immobility was considered as sleep (bottom panels). c Average sleep duration (± s.e.m.) for flies induced to sleep for 24 h by either optogenetic dFB activation (dFB) or Gaboxadol (GAB) compared to controls. All flies were 23E10-Gal4/UAS-Chrimson and exposed to red light, but dFB flies were fed food containing retinal, GAB flies were fed food containing Gaboxadol, while control flies were fed unadulterated food (n = 102 flies for each group). d Arousal thresholds (AT) for 23E10-Gal4/UAS-Chrimson flies exposed to the same conditions as in c. ***p < 0.001, ****p < 0.0001, Kruskal–Wallis test with Dunn’s multiple comparisons test. n = 102 for all groups. Medians (yellow bars) and 75th percentiles (box) and outliers (whiskers) are shown. e Daytime behavioral responsiveness of sleeping flies during recovery following 12 h of nighttime sleep induction or sleep deprivation (SD). Sleep induction methods are as in c and d, or both methods combined (GAB/dFB). *p < 0.05, **p < 0.01, by ANOVA with Tukey’s multiple comparisons. Controls, n = 153; dFB, n = 153; GAB, n = 150; GAB/dFB, n = 102; SD, n = 168. f Daytime behavioral responsiveness of awake flies (see Methods section) during recovery following 12 h of nighttime sleep induction. ****p < 0.0001, by ANOVA with Tukey’s multiple comparisons. The Data are from the same flies as in e. Images: Michael Troup