Abstract
Hippocampal sharp-wave ripples (SWRs) are critical events implicated in memory consolidation, planning, and the reactivation of recent experiences. Under freely moving conditions, a well-established dichotomy exists: hippocampal networks predominantly generate theta oscillations during periods of reward pursuit (preparatory behaviors) and exhibit pronounced SWR activity once the reward is achieved (consummatory behaviors). Here, it was examined how SWRs are modulated by reward delivery and small movements in head-fixed rats. Contrary to the canonical view established in freely moving settings, the results revealed that the dominant and more enduring effect was a sustained suppression of SWR activity immediately following water delivery. Moreover, even minor, localized movements (such as whisking or body adjustments) decreased SWR occurrence, demonstrating that hippocampal ripple generation is highly sensitive to motor engagement, irrespective of reward timing. Such movement-induced suppression of ripples persisted during both sleep-like states and quiet wakefulness, suggesting that while large-scale brain states modulate the overall likelihood of SWR generation, local motor-related influences exert a state-independent inhibitory effect on hippocampal ripples. These results show that SWR modulation by behavioral states and body movements is more context dependent than previously appreciated.
Significance Statement
Small body movements robustly suppress hippocampal sharp-wave ripples across all behavioral states in head-fixed rodents. This reveals a highly context-dependent modulation of SWRs, redefining our understanding of hippocampal ripple dynamics during rest and water consumption.
Introduction
The hippocampus is integral to learning and memory, functioning within a complex network of brain regions including the prefrontal cortex, medial entorhinal cortex, and frontal cortex (Harvey et al., 2023). Among the various electrophysiological phenomena observed in the hippocampus, sharp-wave ripples (SWRs) are prominent. These high-frequency, high-amplitude oscillations occur primarily in the CA3 and CA1 regions and are associated with the reactivation of previously active neural ensembles. This reactivation is thought to contribute to memory consolidation, decision-making, planning, and creative thinking (Ji and Wilson, 2007; Buzsáki, 2015). Moreover, disruptions in SWR activity have been implicated in a range of neurological and psychiatric disorders, including epilepsy, schizophrenia, and Alzheimer's disease (Suh et al., 2013; Zhen et al., 2021).
Extensive research has demonstrated that SWRs occur most frequently during states of immobility, sleep, and consummatory behaviors (Singer and Frank, 2009; Buzsáki, 2015). Consummatory behaviors fulfill basic physiological needs such as eating, drinking, resting, and defecating (Ball and Balthazart, 2008). Historically, SWRs were thought to primarily facilitate learning and memory consolidation once these needs had been met, supporting the retrospective processing of experiences after immediate drives were satisfied.
However, recent investigations have revealed a more nuanced relationship between SWRs and consummatory behaviors. For instance, recent reports demonstrated a decrease in SWR activity during the intake of water rewards (Klee et al., 2021; Sakairi et al., 2025), suggesting that these dynamics may be closely modulated by the animal's current behavioral state. A critical distinction in these studies is whether the animals were head-fixed or allowed to move freely, indicating that SWR patterns can be differentially regulated based on experimental context and the state of motion. These findings underscore the importance of behavioral and environmental factors in shaping hippocampal dynamics.
To further examine how SWRs relate to distinct behavioral states, experiments were conducted in head-fixed rats, with simultaneous tracking of their movements and recording of hippocampal activity. First, the SWR frequency surrounding reward delivery events (randomly provided water) was evaluated, and subsequently, the relationship between changes in the rats’ activity levels and SWR occurrence was assessed throughout the entire session. Through this approach, a more comprehensive understanding of the interplay among SWRs, behavioral states, reward processing, and body movements is provided, offering new insights into how hippocampal neural activity aligns with different behavioral states.
Materials and Methods
Animals and surgery
All experiments were approved by the Animal Care and Use Committee of Institute of Science Tokyo (A2023-116) and were performed in accordance with the Fundamental Guidelines for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions (MEXT, Japan) and the Guidelines for Animal Experimentation in Neuroscience (Japan Neuroscience Society). All surgical procedures were performed under appropriate isoflurane anesthesia, and all efforts were made to minimize suffering (Nonomura et al., 2018; Rios et al., 2023).
Four wild-type Long–Evans rats (8–10 weeks old; 210 ± 35 g; male, 2; female, 2) were kept in their home cages under an inverted light schedule (lights off at 9:00 A.M.; lights on at 9:00 P.M.). The rats were briefly handled by the experimenter (10 min, twice) in advance. For head-plate (CFR-2, Narishige) implantation, the animals were anesthetized with isoflurane (4.5% for induction and 2.0–2.5% for maintenance, Pfizer) using an inhalation anesthesia apparatus (Univentor 400 anesthesia unit, Univentor) and were placed on a stereotaxic frame (SR-10R-HT, Narishige). Lidocaine (AstraZeneca) was administered around the surgical incisions. Reference and ground electrodes (Teflon-coated silver wires, A-M Systems; 125 μm diameter) were implanted above the cerebellum. During anesthesia, body temperature was maintained at 37°C using an animal warmer (BWT-100, Bio Research Center). Analgesics and antibiotics were applied postoperatively, as required (meloxicam, 1 mg/kg, s.c., Boehringer Ingelheim; gentamicin ointment, 0.1%, MSD). A second surgery was performed for later electrophysiological recordings. We made two tiny holes (1.0–1.5 mm in diameter) in the skull and dura mater above the dorsal hippocampi bilaterally (3.8 mm posterior and ±2.0 mm lateral from the bregma). The skull surface and holes were immediately covered with the antibiotic ointment and silicon sealant until the recording experiments.
After full recovery from surgery (3–7 d later), the rats had brief acclimation sessions to the head-fixed apparatus. Specifically, each rat was placed in the recording box and head-fixed for 10 min on two different occasions, mimicking the conditions of the actual recording. This step ensured that the animals gradually adapted to head fixation, minimizing potential stress. The rats had ad libitum access to water during weekends, and during the rest of the week, they obtained water only during the recording session. Additionally, agar was given to the rats in their home cage to maintain them at >85% of original body weight (Isomura et al., 2009; Kimura et al., 2012; Soma et al., 2017).
Electrophysiological recording
Extracellular multichannel recordings of local field potentials (LFPs) and single-neuron spike activity were obtained from the hippocampal CA1 area. In two rats, we simultaneously recorded from the parietal cortex located above the CA1 recording sites (∼1.2 mm below the cortical surface). Two 32-channel silicon probes (ISO-3x-tet-lin with seven tetrode-like arrangements and four vertically aligned channels on three shanks; NeuroNexus Technologies) were inserted vertically into the left CA1 or the parietal cortex, supported by a 2% agarose-HGT (Nacalai Tesque) gel layer on the brain surface.
Recording depth was determined to detect ripple oscillations within the stratum pyramidale (s.p.) of CA1 (Isomura et al., 2006), typically at the second tetrode from the probe tip. Thus, the upper tetrodes were located in the stratum oriens (s.o.) and the lower (tip) tetrodes in the stratum radiatum (s.r.). After the silicon probe was lowered to the appropriate depth, it was left in place for 1 h before initiating the data collection. During this period, the rat remained head-fixed under the same conditions used for recording, allowing additional habituation and minimizing any acute stress response due to handling or probe insertion. In the final recording session, probes were coated with the fluorescent dye DiI (DiIC18(3), PromoKine) to visualize probe tracks post hoc.
Neuronal activity was amplified (32-channel main amplifier: FA-32, Multi Channel Systems; final gain: 2000; bandpass filter: 0.5 Hz–10 kHz) through a 32-channel head-stage preamplifier (MPA 32I, Multi Channel Systems; gain: 10). Signals were digitized at 20 kHz using a 32-channel hard-disk recorder (LX-120, TEAC).
Histology
Following the final recording sessions, rats were deeply anesthetized with urethane (2–3 g/kg, i.p.) and perfused transcardially with cold saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB). The brain was then postfixed and sectioned coronally at 50 µm using a vibratome (VT1000S, Leica). Fluorescent DiI-labeled probe tracks and hippocampal CA1 lamination were examined under a fluorescence microscope (BX51N, Olympus).
SWR, sleep spindles, and LFP delta power analysis
SWR detection followed standard threshold criteria (Csicsvari et al., 2000; Isomura et al., 2006; O'Neill et al., 2006). Briefly, multichannel signals were downsampled to 1 kHz and bandpass filtered in the ripple range (150–250 Hz). The signal envelope was obtained using the Hilbert transform, and SWR events were defined as epochs exceeding 4 standard deviations from the mean (we excluded such events exceeding 9 standard deviations, considering artifacts) for at least 50 ms. The amplitude and peak time of each SWR were defined as the peak amplitude of the enveloped signal in the ripple-band LFP. When SWRs overlapped across multiple channels, the event with the largest peak amplitude was selected for analysis. Artifacts and noise were identified and removed by comparing signals from channels with lower ripple counts and by visual inspection independent of behavior-related information.
For cortical delta power analysis, the LFP signal was bandpass filtered at 0.5–4 Hz and the envelope was estimated via the Hilbert transform. Periods of delta-on activity were defined as those exceeding 4 standard deviations for at least 200 ms (Norman et al., 2021).
Sleep spindles were detected by applying an 8–16 Hz bandpass filter to raw LFP signal. The envelope of the signal was extracted by computing the Hilbert transform of the bandpass-filtered signal. A time window was classified as a spindle if the upper envelope exceeded a threshold of 2.5 SD above the mean once and the lower envelope crossed a threshold of 1.5 SD below the mean for >300 ms (Wang et al., 2024).
Video recording and analysis
Three cameras (Basler ace PoE camera acA1440-73 B/W, Edmund) were positioned at front, diagonal, and overhead angles to record the animals’ movements. The initial frame rate (50 frames/s) was frame blended and interpolated to 30 frames/s for analysis using FFmpeg software (FFmpeg team). Movement detection was based on a pixel-wise absolute difference (PWAD) calculation (Soma et al., 2019):
For every pair of ROIs the Pearson’s correlation coefficient (R) was calculated across the full session. A symmetric distance matrix,
Spike analysis
Offline spike sorting was performed to isolate single units from each tetrode. Spike candidates were initially detected and clustered using EToS (Takekawa et al., 2010, 2012). Manual refinement of these clusters was performed using Klusters and NeuroScope (Hazan et al., 2006), ensuring that single-neuron clusters had a clear refractory period (>2 ms) in autocorrelograms and no refractory period contamination in cross-correlograms with other clusters. Single-neuron clusters were included if they contained at least 250 spikes in total (Kawabata et al., 2020; Mitani et al., 2022).
Experimental design and statistical analyses
Electrophysiological CA1 data were obtained from 66 recording sessions across four rats. In a subset of 12 sessions from two rats, we additionally recorded the activity from deep layers of the parietal cortex to evaluate LFP delta power. Each session lasted ∼90 min, during which water (for the sake of simplicity we will call it reward) was delivered through a spout in front of the animal's mouth at a random time among a uniform distribution from 140 to 220 s, with no external cues. Each delivery consisted of three 5 µl drops administered by a micropump. SWR activity was aligned to both water delivery events and ROI-defined movement events to construct perievent time histograms (PETHs).
SWR frequency was normalized (z-score) on a session-by-session basis using a baseline window (−5 to −2.5 s relative to the event). Two test windows were compared against this baseline: 0 to +0.5 s and 0 to +2.5 s relative to water delivery or movement events. Comparisons were conducted on a trial-by-trial basis within and across sessions using Wilcoxon signed-rank test. Correlations between cortical LFP delta power and CA1 SWR activity were assessed with Pearson's correlation, and significance was evaluated using a t test on the correlation coefficients.
To examine the contribution of cortical delta power and ROI-specific movements to hippocampal ripple probability, we implemented an L1-regularized logistic regression analysis using binarized time series from individual recording sessions. SWR occurrence was first binned into 100 ms intervals across the full session duration, with binary indicators (1 = SWR detected, 0 = no SWR) assigned per bin. Two additional binary regressors were computed for each bin: a Delta regressor marked whether a delta event occurred in the preceding 2 s window, and a Movement regressor indicated the presence of ROI-specific movement within the preceding 1 s. An interaction term (Delta × Movement) was included to assess potential synergistic or antagonistic effects. This balanced dataset was used to fit a logistic regression model predicting ripple presence (dependent variable) from the three predictors (Delta, Movement, Delta × Movement), with a constant intercept term. L1 regularization (lasso penalty) was applied to encourage sparse solutions and reduce overfitting (α = 0.1). Regression coefficients for each predictor were extracted and interpreted across sessions to quantify the direction and consistency of ripple modulation by delta oscillations and ROI-specific movements.
To account for repeated measurements and avoid pseudoreplication, we quantified the movement-related change in ripple incidence for every recording subsession as follows: Δ = SWRTest − SWRBaseline. The analysis set comprised 66 subsessions obtained from four rats (median, 16.5 sessions; range, 10–21). A linear mixed-effects model, Δij = β0 + u0j + εij, was then fitted with a fixed intercept β0 (population mean Δ) and a random intercept u0j capturing between-rat variability [u0j ∼ N (0, σ2rat)]. Models were estimated via restricted maximum likelihood. Significance of β0 was evaluated with a two-tailed Wald t test (α = 0.05); a significantly negative β0 indicates robust suppression of SWR incidence during movement after accounting for inter-animal variability (Pinheiro and Bates, 2000; Lazic, 2010).
All data are expressed as mean ± SD, and n indicates the sample size. Line plots and shaded areas represent mean ± SEM. All statistical analyses were performed with MATLAB's Statistics and Machine Learning Toolbox (The MathWorks) and Python (NumPy 1.26.4, pandas 2.2.2, SciPy 1.12.0, statsmodels 0.15.0). Differences were considered significant at p < 0.05. Blinding and randomization were not performed.
Results
Behavioral states and movement quantification during head fixation
We recorded hippocampal CA1 LFPs together with high-resolution video from three camera angles in head-fixed rats to assess behavioral states and small movements during reward delivery (Fig. 1A, Movie 1). The placement of the recording electrode and the camera orientations allowed precise tracking of orofacial and body movements under near-immobile conditions, ensuring that detected movements were not confounded by locomotion.
Experimental setup, behavioral-state classification, and movement quantification. A, Schematic illustration of the head-fixed experimental configuration. A rat is shown with a recording electrode targeting the hippocampal CA1 region, while three cameras (side, front, top) simultaneously capture the animal's behavior. Water rewards are delivered via a spout positioned near the rat's mouth. B, Representative timeline of a recording session showing the full duration with periods of quiet, reward consumption, and putative sleep (green shading, “Delta on”) identified by cortical LFP delta power (green trace). The blue drops above the timeline mark random water-reward deliveries. The black trace illustrates a representative movement signal from one region of interest (ROI 4: whiskers side), with the red dotted line indicating the threshold used for movement-event detection. The yellow segment highlights the magnified window shown above. C, Example frames from each camera angle (side, front, top) used to define movement ROIs. Left panels show raw video images; right panels display heat maps of pixel-wise differences highlighting areas of increased movement. Labeled ROIs are as follows: ROI 1, tongue; ROI 2, jaw; ROI 3, nose; ROI 4, whiskers (side); ROI 5, whiskers (front); ROI 6, forelimbs; ROI 7, trunk; and ROI 8, tail. D, Representative traces of pixel-wise difference amplitude (PWDA) for each ROI, illustrating how thresholding (red dashed line) is applied to detect movement events. E, Top, Color maps of normalized movement event frequency aligned to reward delivery (time zero) for each ROI, with each horizontal row representing data from a single rat. Bottom, Average perievent time histograms (PETHs) for each ROI, depicting the normalized movement frequency across rats. Shaded areas indicate SEM, and vertical dashed lines mark the time of reward delivery. Please see Extended Data Figure 1-1 for additional data.
Figure 1-1
Within- and across-session progression of movement events. (A) Representative data from three rats illustrating the time course of movement events for each region of interest (ROI) within a single session. Each panel shows the event count over time for one ROI, allowing visualization of how movement frequency changes throughout the session. (B) Example progression of average movement event frequency (events/s) across multiple recording sessions for a single rat. Each trace corresponds to one ROI, revealing the relative stability or variability of movement patterns as the sessions progress. Download Figure 1-1, TIF file.
Quiet resting. Twenty-four-second recording centered on a single water-reward delivery (time 0 s). The top left composite displays simultaneous side-, front-, and top-camera views of the head-fixed rat. The bottom left panel shows the hippocampal LFP filtered at 150–250 Hz in real time. Right panels plot z-scored movement-amplitude traces for the tongue, jaw, nose, whiskers front, and forelimbs ROIs. [View online]
Behavioral epochs were defined based on both experimental events and physiological signals. Specifically, water-reward delivery events were interleaved with intervals of relative immobility, and fluctuations in cortical LFP delta power were used to classify distinct arousal states (Fig. 1B). Epochs characterized by high delta power were regarded as putative sleep, whereas periods of low delta power corresponded to quiet wakefulness and active reward consumption (Fig. 4). Although our analysis is not based on discrete trials, we quantified the continuous transitions between these epochs across 66 sessions from four rats.
To assess the animals’ movements in detail, pixel-wise absolute difference (PWAD) analyses was performed on the video frames (Fig. 1C; see Materials and Methods). ROIs were defined for various body parts, including the tongue, jaw, nose, whiskers, forelimbs, trunk, and tail, capturing reward-locked changes in movement (Fig. 1D). Across all three camera views, these analyses revealed pronounced orofacial and forelimb movements tightly coupled to water-reward delivery, as well as smaller adjustments in body posture and whisking behavior during quiet periods or transitions into rest or sleep states.
Quantitative analysis (Fig. 1E) confirmed that the frequency of orofacial movements (e.g., tongue, jaw, and nose) increased sharply around reward events (tongue, 9.15 ± 1.23 Hz; jaw, 8.45 ± 1.79 Hz; nose, 1.47 ± 0.97 Hz; whiskers side, 1.45 ± 0.82 Hz; whiskers lat, 1.35 ± 0.80 Hz), reflecting the animals’ anticipatory and consummatory responses. In contrast, movements of other body parts (trunk, 0.83 ± 0.54 Hz; tail, 0.71 ± 0.55 Hz) exhibited distinct temporal profiles, often peaking at different times or showing more gradual changes over the trial epoch. Movements remain constant along individual and across sessions (Extended Data Fig. 1-1). Together, the consistency of movements across animals around reward-related events and ROI-based analyses provides a detailed behavioral framework against which we will interpret the temporal modulation of SWR activity in subsequent sections.
Modulation of SWR activity during consummatory behavior
Next, it was examined how hippocampal SWR frequency and associated CA1 neuronal activity were modulated by reward delivery under head-fixed conditions. To do this, LFPs were recorded from the CA1 region and aligned identified SWR events (Fig. 2A) to the timing of water-reward administration (Fig. 2B).
Modulation of hippocampal SWR activity and CA1 spiking by reward events. A, Representative example of SWR and delta-related epoch detection. Top, Raw CA1 LFP trace with two magnified segments (green and red boxes) highlighting transient high-frequency ripple oscillations characteristic of SWRs. Middle, Bandpass-filtered signals (500–3,000 and 150–250 Hz) and their Hilbert transform envelopes. The red dashed line indicates the SWR detection threshold (4 standard deviations). Bottom, Representative cortical 0.5–4 Hz trace and its Hilbert transform envelope used to identify delta-on and delta-off periods. B, Single-session example showing SWR occurrences aligned to reward delivery. Top, Raster plot of SWR events from multiple trials. Bottom, PETH of SWR frequency, illustrating a phasic decrease in SWR activity immediately following the reward (dashed vertical line). C, Population-level SWR activity aligned to reward delivery. Top, Heat map of z-scored SWR frequency, with each row representing one session. The white dotted lines divide sessions by rat, indicated by the color bar on the right. Bottom, Mean population PETH (gray trace) with shaded SEM, overlaid with individual rat PETHs in color. Blue shading represents baseline, and the green and red shaded regions highlight short-latency (0–0.5 s) and extended (0–2.5 s) postreward windows, respectively. D, Representative session showing trial-by-trial changes in SWR rate from baseline to the short-latency (left) and extended (right) postreward periods. Dots represent individual trials; colored dots indicate the mean. Asterisks denote significant differences (*p < 0.05; **p < 0.01; Wilcoxon signed-rank test). E, Population-level comparisons of SWR frequency changes in short-latency (left) and extended (right) postreward intervals relative to baseline. Each rat is shown in a different color (**p < 0.01; Wilcoxon signed-rank test). F, Cross-correlograms of single-unit firing relative to SWR maxima, showing all recorded neurons (top) and the average (bottom). G, Population spike firing rates of CA1 neurons aligned to reward delivery (time zero). The top panel shows the color-coded spike frequency across units, while the bottom panel displays the average firing rate change. Vertical dashed lines denote the time of reward delivery.
Water delivery produced a distinct change in ripple dynamics (Fig. 2B,C). Although some sessions showed a transient increase in SWR frequency immediately after reward delivery (Fig. 2D; p = 0.018 for 0–0.5 s; p = 1.52 × 10−4 for 0–2.5 s; Wilcoxon signed-rank test), the dominant response across the population was a prolonged suppression of SWRs following the initial response (Fig. 2E; p = 2.01 × 10−5; Wilcoxon signed-rank test for 0–2.5 s after reward; Table 1). To address the concern of pseudoreplication and confirm that the overall effect does not stem from a single outlier animal, we next fit a linear mixed-effects model (see Materials and Methods). The model yielded a significantly negative intercept (β = −0.775, p < 0.001), indicating that, on average, the overall suppression effect remains robust once we properly account for repeated measures across subjects. This pattern suggests that in a head-fixed context, the network's response to reward differs from the classic enhancement observed during consummatory behavior in freely moving animals.
Statistical summary
To validate our SWR detection approach and assess the functional relevance of these events, their relationship with neuronal firing was examined (Adaikkan et al., 2024). An increase in the multi-unit activity (MUA, 500–3,000 Hz band filtered signal) related to a high ripple-band power was observed (Fig. 2A). Additionally, using a subset of recordings (n = 12 sessions) where single-unit activity was simultaneously acquired, cross-correlograms were constructed between neuronal firing times and SWR peaks (Fig. 2F). This analysis revealed a significant increase in spike discharge coincident with SWR events, confirming that the detected ripples corresponded to genuine hippocampal network activity.
Finally, changes in spike firing rates of CA1 neurons were examined around reward delivery (Fig. 2G). Consistent with the notion that reward information can transiently enhance hippocampal excitability (Singer and Frank, 2009), most neurons showed increased firing shortly after reward. Nonetheless, a subset of neurons exhibited no significant change or even a reduction in firing, highlighting the heterogeneity of hippocampal network responses to reward events (Sakairi et al., 2025).
These results demonstrate that reward delivery under head-fixed conditions modulates hippocampal SWR activity, with a complex temporal pattern characterized by transient increases in some sessions but a more robust, long-lasting suppression observed across the population.
Modulation of SWRs by orofacial and body movements
Next, the relationship between SWR occurrence and the onset of movement events associated with different body regions was investigated. To avoid confounding influences of reward-related effects on ripple activity, all movement events occurring within ±12 s of reward delivery were excluded. With this constraint in place, SWR activity was aligned to the onset of detected movements in specific ROIs, including the tongue, jaw, nose, whiskers (lateral and front), forelimbs, trunk, and tail. SWR frequencies in a baseline window preceding the movement event were compared with a test window immediately following it.
Orofacial movements led to a predominant decrease in SWR frequency across sessions (Fig. 3A–E; Table 1). Although a minority of sessions showed increased or unchanged SWR activity, the overall trend was a significant reduction relative to baseline levels. This reduction was similarly observed for movements of other body parts, including forelimbs, trunk, and tail (Fig. 3F–H). Although session-to-session variability existed (with some sessions showing increases or no change), the dominant population-level pattern across all these ROIs was a reduction in SWR activity associated with movement events. In each case, the population PETHs revealed a downward shift in SWR frequency following the onset of the corresponding body part movement.
Modulation of SWR activity by orofacial and body movements, independent of reward proximity. A, Top left, Population-level heat map of SWR activity aligned to the onset of tongue movements (top) and the corresponding PETH of SWR frequency (bottom). In the PETH plots, the population average is shown in gray with the SEM shaded, and each rat's average PETH is displayed in a different color. Blue shading indicates the baseline (Pre) window, and light red shading denotes the post-movement (Post) window. On the right of each panel, dot plots compare SWR activity between these two windows for each session; filled symbols indicate sessions with significant changes. B, SWR activity aligned to jaw movements. C, SWR activity aligned to nose movements. D, SWR activity aligned to whisker lateral movements. E, SWR activity aligned to whisker front movements. F, SWR activity aligned to forelimb movements. G, SWR activity aligned to trunk movements. H, SWR activity aligned to tail movements (*p < 0.05; **p < 0.01; Wilcoxon signed-rank test).
Together, these findings indicate that hippocampal SWRs are generally suppressed around the time of active orofacial and body movements, even when excluding movements that occur near reward delivery. This suggests that hippocampal network states conducive to SWR generation are less prominent when the animal engages in even minor, localized movements. This relationship was consistent across multiple ROIs, underscoring the robust link between behavioral activity and hippocampal ripple dynamics.
Delta power correlates with SWR band activity and influences movement-related modulations
To further explore the interplay between behavioral states and SWRs, it was examined how fluctuations in cortical LFP patterns, i.e., delta power (0.5–4 Hz) and sleep spindles, correlate with hippocampal ripple activity (150–250 Hz). Across the 12 sessions obtained from two rats, moment-by-moment fluctuations in the cortical delta envelope provided a reliable index of hippocampal ripple probability. Elevated delta power was accompanied by higher SWR rates (Fig. 4A,B,E). Delta peaks also coincided with sleep spindle events, which themselves correlated with ripples (Fig. 4C,D,F,G). These data position delta oscillations and sleep spindles (associated with low arousal and sleep pressure) as a broad, brain-state gate that raises the baseline excitability of the hippocampal network. Also, the regression between sleep spindles and delta events exhibited an intercept close to zero, implying that spindle events are minimal or nearly absent in the absence of delta events (Fig. 4F). In contrast, both delta–SWR (Fig. 4E) and spindle–SWR (Fig. 4G) correlations showed distinctly positive intercepts, indicating that SWRs occur at low but detectable rates even when spindle or delta events are scarce. This synchronous modulation aligns with previous studies linking large-scale brain states, such as sleep or quiescent wakefulness, to coordinated changes in hippocampal ripple generation (Sirota et al., 2003; Mölle et al., 2006; Goldstein et al., 2018).
Delta-dependent gating of hippocampal sharp-wave ripples and dissociation from movement-related suppression. A, Band-specific LFP traces. Representative 600 s excerpt of a single recording session showing 10–16 Hz filtered-band (magenta), delta-band (0.5–5 Hz, blue), and ripple-band (150–300 Hz, red) signals. The inset highlights detected spindle events (magenta, tick marks) over the raw LFP. B, Delta events versus ripple frequency correlation. Ten-second binned scatterplot from the same session relating delta-event rate to SWR rate. Red line, Least-squares fit; Pearson’s r = 0.463, p = 1.31 × 10−39. C, Spindle events versus delta events correlation. Scatterplot and linear fit (red) for spindle-event rate against delta-event rate (r = 0.591, p = 3.46 × 10−69). D, Spindle events versus ripple frequency correlation. Relationship between spindle-event rate and SWR rate in the same session (r = 0.488, p = 2.15 × 10−44). E, Population delta–ripple correlation. Lines represent individual sessions (n = 12); inset gives the mean Fisher-z-transformed r values (mean = 0.23) and one-sample t test versus 0 (p = 0.0042). F, Population spindle–delta correlation. Same format as E; mean r = 0.47, p = 0.0001. G, Population spindle–ripple correlation. Same format as E; mean r = 0.23, p = 0.0042. H, Logistic-regression coefficients for individual ROI movements. L1-regularized logistic regression predicting SWR probability from single-ROI movement events. Error bars show coefficients from each session. I, Logistic-regression coefficients for delta power. Corresponding positive coefficients for delta-band power across sessions. J, Interaction coefficients (delta × movement). Interaction-term estimates for each ROI; most centered on zero, indicating little synergistic or antagonistic modulation beyond the main effects. K, Hierarchical clustering of ROI movements. Dendrogram constructed from 1—correlation distances of movement time series, revealing clusters of correlated motor motifs (e.g., trunk–tail). L, Logistic-regression coefficients after grouping ROIs. Coefficients recomputed with predictors collapsed into Licking, Orofacial, Whisking, Forelimbs, and Body movements categories. Delta remains significantly positive, grouped movements remain negative, and only the Delta × Body interaction reaches significance (p < 0.05). M, Movement amplitude during delta states. Box-and-whisker plots comparing mean pixel-difference amplitude for five grouped motor categories during delta-on (red) versus delta-off (blue) epochs. Dots, individual sessions (Wilcoxon signed-rank). N, Movement frequency during delta states. Mean event rate (Hz) for the same categories and statistical testing as in H. Error bars, SEM across sessions; *p < 0.05; **p < 0.01. Please see Extended Data Figures 4-1, 4-2 for additional data.
Figure 4-1
Impact of cortical delta power on movement-related hippocampal SWR modulation across behavioral states. (A) Movement-related modulation of SWRs during delta-on (sleep-like) epochs. For each ROI (tongue, jaw, nose, whiskers lateral, whiskers front, forelimbs, trunk, tail), the left panels show the average PETHs of SWR frequency aligned to movement events. Blue and red shaded areas represent the baseline and test windows, respectively. The right panels illustrate session-by-session comparisons of pre- vs. post-movement SWR activity (blue and red dots, respectively), with filled symbols indicating sessions that reached statistical significance. (B) The same analysis as in (A), but during delta-off (quiet awake) epochs. (* = p < 0.05; ** = p < 0.01; Wilcoxon signed-rank test). Download Figure 4-1, TIF file.
Figure 4-2
(A) Session-by-session matrix illustrating changes in SWR frequency aligned to the onset of movements for multiple body parts during delta-off epochs (tongue, jaw, nose, whiskers side, whiskers front, forelimbs, trunk, tail). Each row represents one recording session, and each column corresponds to a specific body part. Colors indicate the normalized change in SWR activity relative to baseline: blue denotes a decrease, and red denotes an increase. Statistical significance, assessed by Wilcoxon signed-rank test, is indicated by asterisks (*p < 0.05; **p < 0.01; Wilcoxon signed-rank test). (B) Same as (A) for movements for multiple body parts during delta-on epochs. Download Figure 4-2, TIF file.
We next asked whether self-generated behavior could override this permissive influence. L1-regularized logistic models that treated each video-derived motor feature, delta events, and their interaction as simultaneous predictors revealed a consistent pattern: every movement carried a negative weight whereas delta remained strongly positive (Fig. 4H,I). Interaction terms were near zero, indicating that movement suppression and delta facilitation add linearly rather than synergistically (Fig. 4J). Separating perimovement analyses by brain state revealed an identical qualitative outcome. In putative sleep-like (delta-on) epochs, ripple rates dropped the moment a movement began; the same pattern was evident, albeit slightly weaker, during delta-off quiet wakefulness (Table 1; Extended Data Fig. 4-1). Taken together, these findings demonstrate that (1) cortical delta oscillations promote hippocampal ripples and (2) the suppressive influence of movement on ripples is remarkably robust, though the degree of modulation sometimes varied across sessions, surviving even within delta-dominated, putative sleep-like periods. While the overarching pattern of movement-induced SWR suppression remained consistent across delta-on and delta-off conditions (Table 1; Extended Data Fig. 4-2), subtle differences in the magnitude or significance of SWR modulation suggest that the underlying brain state, as reflected by delta power, can influence the extent to which hippocampal ripples are affected by behavioral events.
Because several ROIs captured partially overlapping behaviors, we next examined their redundancy. Hierarchical clustering of the eight-movement time series based on 1 − R revealed two principal clusters: (1) tongue, jaw, nose, and whisker motions (orofacial) and (2) trunk and tail displacements accompanied by forelimb retractions (body-centered; Fig. 4K). Highly correlated ROIs were further consolidated into five broad categories (Licking, Orofacial, Whisking, Forelimbs, Body). Refitting the model with these five movement groups produced similar results (Fig. 4L): delta remained a strong positive predictor (p < 0.01), all grouped movements were negative (p < 0.01 for four of the five), and only the delta × Body interaction attained significance (p = 0.03), indicating a mildly weaker suppression of ripples by gross body motions during high-delta periods. Although delta events coincided with visibly quieter behavior, animals were not completely still. Both the amplitude and the frequency of detected movements fell during delta-on windows, yet small movements persisted (Fig. 4M,N). These combined behavioral and electrophysiological markers showed that the animal can exhibit movements and enter a sleep-like state despite being head fixed (Movies 2, 3; Tiriac et al., 2012; Dos Santos et al., 2019).
Brief adjustments. Twenty-four-second recording aligned to water-reward delivery (time 0 s). Layout is identical to Movie 1. Camera views capture short forelimb and orofacial movements. Corresponding transient peaks appear in the forelimb and orofacial ROI traces, whereas other ROIs remain near baseline. Ripple-band LFP is displayed continuously. [View online]
Prolonged immobility. Twenty-four-second recording aligned to water-reward delivery (time 0 s). Layout is identical to Movie 1. Camera views show the rat remaining largely motionless for the duration of the clip. All ROI traces remain at baseline except for occasional minor deflections. The hippocampal 150–250 Hz LFP is displayed in the bottom left panel. [View online]
In summary, these results demonstrate that SWR activity is tightly coupled to cortical delta oscillations and that this relationship remains evident when examining the effects of movement on SWRs across different brain states. Whether during delta-on (sleep-like) or delta-off (awake quiet) epochs, movement-related suppression of SWRs persists, being consistent across body parts (Fig. 5), underscoring the multifaceted and state-dependent regulation of hippocampal network dynamics.
Summary of session-wise impact of individual and grouped movements on ripple likelihood. Schematic summary of the SWR activity related to body movements. Each body part of the rat is paired with a donut plot indicating the percentage of sessions that showed significant decreases (blue), significant increases (red), or no significant change (gray) in SWR activity. The outer donuts summarize the same analysis after pooling ROIs into five movement clusters (Licking, Whisking, Orofacial, Forelimbs, and Body). Colors and numeric labels follow the same convention as the inner rings.
Discussion
In this study, we investigated how hippocampal SWRs are modulated by behavioral states, reward delivery, and localized movements in head-fixed rats. Building on previous work that established SWRs as critical for memory consolidation, decision-making, and planning (Ji and Wilson, 2007; Girardeau and Zugaro, 2011; Buzsáki, 2015) and that their disruption can lead to pathological outcomes (Suh et al., 2013; Zhen et al., 2021), our findings provide new insights into the dynamic interplay between hippocampal network activity and specific behavioral components.
Traditionally, research conducted in freely moving rodents has outlined a clear dichotomy between preparatory and consummatory behavioral states (O'Keefe and Nadel, 1978; Csicsvari et al., 2007; Buzsáki, 2015). During preparatory periods, when animals actively pursue rewards, hippocampal activity is dominated by theta oscillations, which are associated with spatial processing and attention. In contrast, when animals transition into a consummatory phase (such as consuming a reward), SWRs typically increase, suggesting the facilitation of the consolidation and replay of recently acquired information (Singer and Frank, 2009; Buzsáki, 2015). These distinctions have been observed in tasks where animals are engaged in foraging, navigation, and consumption.
Our findings, however, suggest that under head-fixed conditions, the modulation of SWRs by reward delivery is more nuanced and context dependent. Although some sessions exhibited a brief, transient increase in SWR frequency immediately after reward delivery (Fig. 2), the predominant pattern was a prolonged suppression of SWRs across the population (Fig. 5). This suppression was observed across a range of movements, including even minimal orofacial and body adjustments, supporting the idea that any motor activity can inhibit hippocampal ripple generation. In a similar direction, during delta-off epochs, regardless of whether some sessions and ROIs display no significant changes or occasionally an increase in SWRs, the dominant pattern remains a suppression of SWR activity around movement events. This finding is consistent with recent observations reporting similar results in head-fixed rodents (Klee et al., 2021, Sakairi et al., 2025). Thus, although hippocampal networks retain the capacity for SWR enhancement under certain conditions, head fixation and the associated constraints on naturalistic movement and behavioral sequences appear to reshape the interplay between reward consumption and ripple generation. The lack of full-bodied locomotion and extensive exploratory behaviors in head-fixed paradigms indicates that in our experimental setup the behavioral state is not directly analogous to the preparatory and consummatory phases seen in freely moving tasks.
Beyond reward modulation, we found that even minor orofacial and body movements strongly suppressed SWR occurrence (Figs. 3, 4). This phenomenon extends the classic view that SWRs occur predominantly during immobility, quiet wakefulness, and sleep (Foster and Wilson, 2006; Buzsáki, 2015). These results reveal that small, localized movements, far from the vigorous locomotion typically studied, are sufficient to disrupt the conditions favoring SWR generation. While in freely moving rodents, preparatory states often involve exploratory locomotion that can suppress ripples (Sullivan et al., 2011; Kay and Frank, 2019), our findings suggest that even in a head-fixed setting, minimal motor engagement still reallocates hippocampal resources away from internal replay and toward processing immediate sensory-motor inputs. Thus, the suppression of SWRs during small movements may represent a fundamental and conserved principle of hippocampal function, one that transcends differences in the environmental complexity or degree of physical freedom available to the animal. A potential mechanism underlying this movement-induced suppression of SWRs is the cholinergic system, particularly inputs from the medial septum. Previous work has shown that optogenetically stimulating septal cholinergic neurons both enhances theta oscillations and suppresses sharp-wave ripples (Vandecasteele et al., 2014) and that in natural conditions, acetylcholine (ACh) levels rise during active behaviors and fall during ripple generation (Zhang et al., 2021). Because cholinergic neuron firing correlates with movement, it is plausible that even small motor activity boosts ACh release, thereby inhibiting the hippocampal network states that support SWR occurrence. Future experiments could test this hypothesis directly.
The analyses also showed that SWR activity is closely coupled with cortical delta oscillations (Fig. 4), a known hallmark of global brain states such as sleep and quiet wakefulness (Sirota et al., 2003; Mölle et al., 2006). Interestingly, spindle events appeared fundamentally dependent on cortical delta states, as evidenced by the near-zero intercept of their correlation (Fig. 4F). In contrast, both delta–SWR and spindle–SWR correlations had distinctly positive intercepts (Fig. 4E,G), suggesting that hippocampal ripples, although strongly facilitated by cortical delta and spindle activity, can nevertheless occur spontaneously, at baseline levels, even when such cortical signals are sparse or absent. This hierarchical organization implies that SWRs represent an intrinsic hippocampal phenomenon, whose expression is modulated, but not exclusively determined, by cortical state fluctuations. Even during high delta and high sleep spindles epochs, putative sleep, rodents are not perfectly still. REM bouts contain rapid limb and whisker twitches, whereas non-REM periods feature brief micro-arousals with small head or body adjustments that momentarily activate forebrain sensorimotor networks (Tiriac et al., 2012; Dos Santos et al., 2019). Whether in these delta-on (sleep-like) or delta-off (quiet-awake) epochs, the fundamental movement-related suppression of SWRs persisted (Fig. 4, Extended Data Fig. 4-1). Thus, while delta power and spindle activity enhance the excitability landscape promoting SWR generation, they do not override movement-related suppression. This interplay between global brain-state signals and local motor-related inputs allows the hippocampus to flexibly balance mnemonic processing with the demands imposed by ongoing behavioral interactions with the environment (Jadhav et al., 2012; Ólafsdóttir et al., 2018).
These results further underscore the complexity of SWR regulation: reward and consummation, in which freely moving studies have often associated with enhanced SWR activity (Singer and Frank, 2009), here yielded a more context-sensitive pattern. The restricted and highly controlled conditions of head fixation likely reduce spatial navigational demands and the variety of preparatory actions that characteristically precede reward consumption in free-ranging foraging scenarios. As a result, hippocampal circuits may shift away from replaying reward-related experiences and instead maintain a lower excitability state after a brief postreward transient. This difference highlights the importance of behavioral context and environmental constraints in interpreting hippocampal activity patterns and their functional significance.
Though our head-fixed preparation allowed precise monitoring of small movements and provided stable electrophysiological recordings, it also limited the behavioral repertoire and spatial exploration of the animals. Future research should replicate these experiments in freely moving settings to confirm the generality of our findings, particularly regarding the dissociation between preparatory and consummatory states (e.g., under more cognitive demanding tasks, requiring a higher engagement in attention, discrimination, or decision-making). Additionally, exploring how SWR modulation by movement and reward differs across hippocampal subregions or species could clarify the universality of these regulatory mechanisms.
In summary, this study provides an alternative view of SWRs as dynamic phenomena influenced not only by global brain states and reward conditions but also by small, localized movements (Fig. 5). By comparing our results with the classical paradigms of freely moving animals, we highlight that SWR modulation is context dependent. This contributes to a more nuanced understanding of hippocampal function, demonstrating that even within similar behavioral states (e.g., consummatory behavior), the neural underpinnings of SWR generation can shift depending on the sensory-motor context and environmental constraints. Such flexibility ensures that the hippocampus can continuously adapt its memory-related activity amid an ever-changing environment.
Footnotes
The authors declare no competing financial interests.
We thank all members of Isomura laboratory, especially Drs. Masanori Kawabata, Riichiro Hira, Yutaka Sakai, and Tomomi Sakairi. This work was supported by Grants-in-Aid for Scientific Research (B) (JP23H02589 to Y.I.), for Transformative Research Areas (A) (JP21H05242 to Y.I.), and for Challenging Research (Exploratory) (JP24K21999 to Y.I.) from MEXT and JSPS; by the Takeda Science Foundation (Y.I.); and by Center for Brain Integration Research, Institute of Science Tokyo.
This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license, which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.
References
In this issue
