Sleep Facilitates Memory by Blocking Dopamine Neuron-Mediated Forgetting

Abstract

Early studies from psychology suggest that sleep facilitates memory retention by stopping ongoing retroactive interference caused by mental activity or external sensory stimuli. Neuroscience research with animal models, on the other hand, suggests that sleep facilitates retention by enhancing memory consolidation. Recently, in Drosophila, the ongoing activity of specific dopamine neurons was shown to regulate the forgetting of olfactory memories. Here, we show this ongoing dopaminergic activity is modulated with behavioral state, increasing robustly with locomotor activity and decreasing with rest. Increasing sleep-drive, with either the sleep-promoting agent Gaboxadol or by genetic stimulation of the neural circuit for sleep, decreases ongoing dopaminergic activity, while enhancing memory retention. Conversely, increasing arousal stimulates ongoing dopaminergic activity and accelerates dopaminergic-based forgetting. Therefore, forgetting is regulated by the behavioral state modulation of dopaminergic-based plasticity. Our findings integrate psychological and neuroscience research on sleep and forgetting.

Figure 1. Ongoing dopamine neuron (DAN) activity is regulated by behavioral state

(A) Tethered fly walking on a ball supported by air during in vivo imaging (see Experimental Procedures for details). (B) Left, DANs (MV1 in blue, and V1 in purple) recorded in this study. Right, representative time averaged image of GCaMP^3.0 signal with regions of interest for both neurons shown in color. (C) Above, recording of ball rotation. Below, co-recorded MV1 and V1 Ca²⁺ activity. Active behavioral time periods shaded in grey. (D) Mean DAN activity in MV1 was robustly higher during active periods, whereas activity in V1 was slightly reduced (*, P < 0.001, n ≥ 8). (E) MV1 activity, but not RFP signal alone or V1 activity, was significantly correlated with ball rotation using a normalized cross-correlation analysis after high frequency filtration (*, P = 0.0078, n = 9). RFP signal was produced by a transgene expressing RFP (Pramatarova et al., 2003) in the DANs to control for motion artifacts. (F) MV1 DAN activity increased with transition into active state (left, * indicates points higher than all points before transition, P < 0.05, n = 105 transitions across 9 animals) and decreased with entry into rest state (right, * indicates points lower than all points before transition, P < 0.05, n = 115 transitions across 9 animals) whereas V1 activity was unaffected by transition (n ≥ 62 transitions across 8 animals). Here and throughout: data are represented as mean ± SEM. See also Figure S1.

Figure 2. Gaboxadol induces sleep and rapidly and reversibly eliminates movement and ongoing DAN activity

(A) Sleep profile for flies fed Gaboxadol (0.1 mg/ml) across one-day pre-treatment (“Pre”), treatment (“Treat”), and post treatment (“Post”) periods, with day-time and night-time periods in white and grey shading, respectively. (B) Gaboxadol increased day-time sleep during treatment and decreased day-time sleep after drug removal in flies fed 0.1 mg/ml Gaboxadol (*, P < 0.0002, n = 30-31). (C) Total activity evoked (see Experimental Procedures) from mechanical stimuli was decreased during treatment in flies fed Gaboxadol (*, significantly different than all other groups that day, P < 0.05, n = 14-16). (D) Fly hanging from an in vivo recording chamber. (E) Gaboxadol administration protocol (“G/S”, Gaboxadol/Saline) used for panels F to H. (F) Representative recordings of movement and MV1 activity during drug perfusion. Mean movement (G) and MV1 activity (H) during the 5-20 min window was decreased by 0.01 and 0.1 mg/ml Gaboxadol perfusion (*, P < 0.05, n = 6). (I) Gaboxadol (0.01 mg/ml) administration and wash-out protocol used for panels J to L. (J) Representative recordings from experiments in I. Mean movement (K) and MV1 activity (L) of flies exposed to 0.01 mg/ml Gaboxadol was decreased during treatment (“Treat”) but recovered fully after washout (“Post”)(*, P < 0.0001, n = 7). See also Figure S2.

Figure 3. Sleep circuit stimulation induces sleep, and rapidly and reversibly eliminates movement and ongoing DAN activity

(A) R23E10-gal4 expression (green, anti-GFP; magenta, anti-nc82) in the dorsal fan shaped body (“dfsb”) sleep circuit. (B and C) Hourly sleep profiles (B) for genotypes (left) during the day prior (“Pre”), during (“Stim”), and after (“Post”) temperature modulation (above). (C) Day-time sleep increased during dfsb stimulation and decreased the next day in the experimental genotype but not the control genotypes (**, P =0.0001, *, P =0.0027, n = 20). (D to H) Activity from neuropil regions of DAN innervation (illustrated in D, “PR”, protocerebrum) and movement were simultaneously recorded (representative traces in G) during temperature modulation (F). The treatment period was divided into 6, 20 min epochs. The treatment window in panel G represents the third [Treat (3)] epoch. (H) Mean movement and relative DAN activity from flies (genotypes, left) across time, with the data from each of the six epochs during the treatment period plotted between the pre- and post-treatment time windows. Movement and activity in MV1 and PR decreased during dfsb stimulation and returned to “Pre” levels afterwards (*, P < 0.05, n = 10-16). Colored lines represent temperatures (light, 23°C; dark, 34°C). See also Figure S3.

Figure 4. Increasing sleep after learning enhances memory retention

(A and B) Starved flies fed either 0.1 mg/ml Gaboxadol for 1 hr (“Gab→Ctrl”) or continuously (“Gab→Gab”) exhibited increased sleep by 20 min (A, *, P < 0.05, n ≥ 15) and increased total sleep during first 8 hr (B, *, P < 0.05, n ≥ 15), compared to starved flies fed control food (“Ctrl→Ctrl”). (C and D), Flies fed 0.1 mg/ml Gaboxadol according to experimental timeline illustrated at the top had enhanced 6 (C, *, P ≤ 0.005, n = 16) and 8 hr (D, *, P ≤ 0.015, n = 12) memory retention. (A=Acquisition, G/C=Gaboxadol or Control food, R=Retrieval). (E to L) Flies (genotypes to the left) with sleep circuit stimulation just after learning (timeline in E) exhibited increased total sleep during stimulation prior to the 3 (F) and 6 hr (G) retrieval time points (sum in H, *, P < 0.0001, n = 16), and enhanced 3 and 6 hr memory retention (I, *, P ≤ 0.0213, n = 8). (J) Flies with simultaneous sleep circuit and c150-gal4 DAN stimulation had increased sleep (J, sum in K, * P < 0.0001, n = 20) but markedly reduced 3 hr memory retention (L) (*, P < 0.0009, n = 8). See also Figure S4.

Figure 5. DANs respond strongly to mechanical stimuli

(A) Mean movement and neuronal activity within MV1, PR, and V1 DAN neuropil before (“Pre”), after (“Post”), and during six airpuff stimuli (“Airpuff”, arrows). Light color lines indicate SEM. (B) All three DAN regions responded to all airpuffs (significantly different from zero, P < 0.0005, n = 12-14). Responsiveness in PR and V1 decreased gradually with repeated airpuffs (slope significantly different from zero, *, P < 0.0154, n = 12-14). Mean movement (C) and ongoing MV1 activity (D) increased after airpuff stimulation (*, P ≤ 0.0295, n = 12-14).

Figure 6. Mechanical stimuli increase population activity after learning and induce forgetting through the DAN forgetting pathway

(A to C) Flies mechanically stimulated (“Stim”) with a shaking platform (protocol and illustration in A) after learning had increased population activity during stimulation (B) leading to a robust increase in mean population activity (C) across the treatment window (“Treat”) but not after (“Post”) compared to un-stimulated flies (“Ctrl”)(*, P =0.0002, n = 8). (D) Wild-type Canton-S flies exhibited decreased 3hr memory retention after 80 min of mechanical stimulation just after learning (*, P < 0.0013, n = 8). (E) Learning and immediate memory retrieval were not disrupted by mechanical stimulation given prior to learning. (F) Mechanical stimulation at either 23°C (left) or 32°C (right) decreased 3 hr memory retention except when synaptic output from c150-gal4 DANs was blocked (*, P < 0.05, n = 8). A gal80 transgene expressed from a mushroom body promoter (MBgal80; Krashes et al., 2007) was employed to block any gal4-regulated transcription in the mushroom bodies. “ns”, not significant. See also Figure S5.

Figure 7. Behavioral state controls memory dynamics via DANs

Animals continuously shift behavioral states between rest and arousal. Internal drive or salient external stimuli, such as the unconditioned stimulus (US), push an animal into an aroused and active state, which in turn activates DANs that innervate the MB memory center. Activity of the DANs has been proposed to convey the US required to be integrated with the odor CS, thus altering the connection between MB neurons and output neurons that drive behavior. We propose that arousal-induced DAN activation increases plasticity in the MB neurons to facilitate memory updating. Memory updating can either represent new learning, when this DAN plasticity is temporally coincident with an associated conditioned stimulus (CS), like odor, or the forgetting of a previously existing odor memory when the same odor that formed the memory is no longer coincident with DAN activation. Sleep benefits memory retention by shifting the behavioral state away from arousal, thus decreasing DAN mediated plasticity and updating. In addition, stable rest or the state of sleep, with the associated reduced DAN activity, may be required for further enhancement of memory through consolidation.