Brain oscillations are associated with many cognitive functions (Buzsáki et al., 2013; Colgin, 2016; Pesaran et al., 2018) and are thought to reflect complex interactions of neural activity from diverse populations of interconnected neurons (Pesaran et al., 2018). For instance, it is well established that the hippocampus is critical for spatial learning and memory (O’Keefe and Nadel, 1978) and that distinct oscillatory states are observed in hippocampal subregion CA1 during spatial navigation in rodents (see Colgin, 2016). CA1 theta power is prominent during spatial exploration, increases with running speed, and its phase is thought to link representations of spatial locations within a trajectory (Buzsáki and Moser, 2013; Drieu and Zugaro, 2019). CA1 also exhibits transient increases in slow (25–55 Hz) and fast gamma (60–100 Hz) power during exploration, which are thought to be associated with retrieval and encoding processes, respectively (Colgin et al., 2009). It remains unclear, however, the degree to which the aforementioned frequency content and associated functions observed during spatial exploration generalize to nonspatial forms of information processing.

In addition to spatial information, accumulating evidence indicates that the hippocampus plays a key role in the processing of temporal information. Consistent with its unique architecture and connectivity (McNaughton and Morris, 1987; Lisman, 1999; Foster and Knierim, 2012; Buzsáki and Tingley, 2018), a growing literature shows that the hippocampus is critical for remembering sequences of nonspatial events (Fortin et al., 2002; Kesner et al., 2002; Allen and Fortin, 2013; Eichenbaum, 2014) and hippocampal neurons code for temporal relationships among such events (MacDonald et al., 2011; Allen et al., 2016; Shahbaba et al., 2022). However, little is known about the oscillatory dynamics associated with this fundamental type of information processing in the hippocampus. Previous studies have observed theta oscillations in CA1 while rodents sampled task-relevant nonspatial stimuli, though the frequency range tended to be lower than during running. For example, ~4–8 Hz oscillations are observed in tasks using olfactory stimuli (Martin et al., 2007; Igarashi et al., 2014; Allen et al., 2016). In addition, the hippocampus exhibits oscillations in the 20–40 Hz range during the processing of odor information (Martin et al., 2007; Kay, 2014; Igarashi et al., 2014; Allen et al., 2016), with this signal varying in power across the proximodistal axis of CA1 (higher power in distal than proximal CA1; Igarashi et al., 2014). Notably, the 20–40 Hz frequency range overlaps with slow gamma (~25–55 Hz) as well as with beta (~15–35 Hz) and its functional relevance remains poorly understood. Based on findings from striatal and cortical circuits (Leventhal et al., 2012; Engel and Fries, 2010; Schmidt et al., 2019), beta has been hypothesized to serve a ‘gating’ function and reflects a ‘status quo’ signal bridging the period between a decision and a response. It is unclear whether this function extends to the role of 20–40 Hz in hippocampus and, more generally, how the recruited spectral profiles in CA1 support memory for temporal order.

To address these issues, we examined CA1 local field potential (LFP) recordings while rats performed a hippocampus-dependent, odor-based sequence memory task (Fortin et al., 2016). This task offers precise time-locking to stimulus presentations and responses, as well as different trial types associated with distinct cognitive demands. As in our previous work (Allen et al., 2016), we report power increases in the 20–40 Hz and low-theta frequency ranges during the odor sequence processing periods. Here, we extend these results by demonstrating that the same electrodes exhibited distinct spectral content in a different state (running on the track), which was characterized by high-theta power and a broadband increase in power for frequencies above 24 Hz. Notably, power in the recruited frequencies varied along the proximodistal axis of CA1 in a behavioral state-dependent manner. Power in the 20–40 Hz range increased with session performance, whereas power in other frequency ranges did not. This performance effect was primarily driven by activity on trials involving a sustained hold response. More specifically, we found that 20–40 Hz power gradually increased during odor presentations and reached a plateau (steady state) on these ‘hold-response’ trials, and that this steady-state power in the proximal segment of CA1 was key to predicting trial accuracy. Collectively, these results suggest that 20–40 Hz power dynamics in the proximal segment of CA1 are linked to processing the temporal context of nonspatial events and reflect a post-decision state that may serve to protect local, task-critical information processing from disruption and interference until a response can be performed.

Five rats were trained on an odor-based sequence memory task that critically depends on the hippocampus (Fortin et al., 2016). In this task, animals receive repeated presentations of a sequence of five odor items and must correctly identify each item as being presented ‘in sequence’ (InSeq; e.g., ABC…) or ‘out of sequence’ (OutSeq; e.g., ABE…) to receive a water reward (Figure 1A and B). More specifically, each odor presentation is triggered by the rat’s nose entering the odor port and the duration of the nosepoke response is used to behaviorally indicate whether the presented item is thought to be InSeq (by continuing to hold until an auditory signal occurs at 1.2 s) or OutSeq (by withdrawing before the signal). Note that the exact time cutoff used to classify InSeq and OutSeq responses was determined separately per animal based on the intersection of their distributions of nosepoke durations on InSeq and OutSeq trials (Figure 1G; ~1.1 s). This is to account for the observation that animals occasionally withdrew just before the 1.2 s threshold (because of the predictable timing of the auditory stimulus), suggesting that these are anticipatory errors not incorrect InSeq/OutSeq decisions. Importantly, this slight adjustment in cutoffs (~1.1 s instead of 1.2 s) more accurately reflects the animals’ decisions but does not change the pattern of results observed. For InSeq items, hold durations longer than the identified time cutoff were categorized as correct responses (InSeq+) and shorter ones as incorrect responses (InSeq-; Figure 1C). The converse was true for OutSeq items: hold durations shorter than the time cutoff were categorized as correct responses (OutSeq+) and longer ones as incorrect responses (OutSeq-).

Figure 1

Download asset Open asset

Following a training period of 2–4 months, animals were surgically implanted with a microdrive over the dorsal CA1 region of the hippocamps and tetrodes were then slowly driven to the pyramidal layer over a period of 2–3 weeks. LFP activity was then recorded from tetrodes spanning much of the CA1 proximodistal axis as animals performed the task (Figure 1D–F). We analyzed LFP activity from three sessions: a session in which animals were tested on the sequence learned preoperatively (ABCDE; well-trained session; n = 5 rats), as well as two successive sessions testing a novel sequence (VWXYZ; novel 1 and novel 2 sessions; n = 4 rats, as one animal damaged its microdrive after the well-trained session). According to our previously established performance metric (sequence memory index [SMI]; Allen et al., 2014), behavioral accuracy in these sessions ranged from low (novel 1; SMI = 0.0619 ± 0.0664; mean ± SEM), to moderate (novel 2; SMI = 0.2624 ± 0.0671), to high (well-trained; SMI = 0.5906 ± 0.0578), providing an opportunity to compare LFP activity across three different levels of performance. Note that the SMI ranges from 0 to 1 and takes into account the uneven distribution of trial types (~90% of trials are InSeq), which makes it a better indicator of performance than percent correct. For additional perspective, the SMI of 0.59 observed in the well-trained session represents asymptotic task performance that corresponds to 92.54 and 68.09% correct on InSeq and OutSeq trials, respectively. In contrast, performance in the novel 1 session is not different from chance (SMI of 0), which is defined by a response pattern following the probabilities of each trial type (in this case, ~90% of trials were InSeq and subjects were 88.25 and 17.47% correct on InSeq and OutSeq trials, respectively). Performance in the novel 2 session is significantly above that of novel 1 but below that of well-trained (92.16 and 31.87% correct on InSeq and OutSeq trials).

Odor sequence processing and running epochs are associated with distinct spectral features, which vary across the CA1 proximodistal axis

We began by investigating whether distinct spectral features in CA1 accompany odor sequence and running periods of the task, and whether the observed spectral content varied across the proximodistal axis of CA1. To do so, group (n = 5) peri-event spectrograms were generated from four electrode locations along the proximodistal axis and aligned to odor processing (Figure 2A) and running (Figure 2B) epochs. For each electrode site, spectrograms were generated using analytic Morlet wavelets and spectral power was z-score normalized to power at the same site during a 30 min period of the recording session. Data were taken from the well-trained session in which animals performed at a high level. For all group analyses, a random intercept linear mixed-effects model (LMEM ANOVA) was fit to the data where electrodes, sessions, or trial types were treated as fixed effects and subjects were treated as a random effect (to account for inherent correlations from repeated measures at the subject level; see Materials and methods for model formulation). For all analyses, data from the same trial type (e.g., InSeq+) was collapsed across sequence positions because no clear modulation of power by position was observed (e.g., Figure 2—figure supplement 1).

Figure 2 with 1 supplement see all

Download asset Open asset

We found that power in the 20–40 Hz and theta (4–12 Hz) range observed during odor presentations showed distinct patterns across the proximodistal axis (Figure 2A, C and D; both frequency ranges defined a priori). More specifically, 20–40 Hz power (110–1200 ms period following port entry) increased toward proximal CA1 (LMEM ANOVA significant effect of electrode: F_3,2214 = 41.995, p<0.001). Power in the most proximal site (electrode 1) showed significantly higher power compared to the two most distal sites (electrodes 3 and 4) (t₂₂₁₄ = –8.072, p<0.001 and t₂₂₁₄ = –8.024, p<0.001, respectively). Power in electrode 1 did not differ significantly from that of electrode 2 (t₂₂₁₄ = –0.206, p=0.837). In contrast, theta power in the same period increased toward distal CA1 (LMEM ANOVA significant effect of electrode: F_3,2214 = 60.952, p<0.001). Electrodes 2–4 showed significantly higher theta power compared to electrode 1 (t₂₂₁₄ = 5.008, p<0.001, t₂₂₁₄ = 11.99, p<0.001, and t₂₂₁₄ = 10.733, p<0.001, respectively).

The same electrodes exhibited a different spectral pattern during running periods, which occurred between sequence presentations (Figure 2B, E and F). More specifically, the 20–40 Hz power observed during odor sampling was weak during running. Additionally, contrary to the odor processing periods, theta power during running was significantly higher in proximal CA1 (LMEM ANOVA significant effect of electrode: F_3,2768 = 71.005, p<0.001; Figure 2F) and was characterized by a higher center frequency (~9 Hz during running compared to ~6 Hz during odor sampling). Electrode 1 showed significantly higher power compared to electrodes 3 and 4 (t₂₇₆₈ = –8.173, p<0.001, and t₂₇₆₈ = –12.414, p<0.001, respectively) while power in electrode 1 did not differ significantly from that of electrode 2 (t₂₇₆₈ = –0.876, p=0.381). The running period was also characterized by increased high-frequency broadband power (>24 Hz to avoid theta’s first harmonic), which significantly increased toward distal CA1 (LMEM ANOVA significant effect of electrode: F_3,2768 = 181.34, p<0.001; Figure 2E). Electrodes 3 (t₂₇₆₈ = 13.255, p<0.001) and 4 (t₂₇₆₈ = 19.53, p<0.001) showed significantly higher power compared to electrode 1. Power in electrode 2 did not differ significantly from that of electrode 1 (t₂₇₆₈ = 1.051, p=0.293).

CA1 20–40 Hz power increases with session performance

We then examined whether power in the 20–40 Hz range was linked to performance and whether this association varied along the proximodistal axis. To do so, we extended the previous analyses, which were applied to the well-trained session, to two consecutive sessions in which animals learned a novel odor sequence (correctly identified InSeq trials only). This allowed us to compare power across three sessions characterized by low (novel 1, first session on novel sequence), moderate (novel 2, second session on novel sequence), and high levels of performance (well-trained session; Figure 3). Unlike the previous analyses, here we evaluated 20–40 Hz power during the 250 ms time window preceding the withdrawal response to capture the period of high power observed toward the end of odor presentations (at which point power has plateaued; Figure 3B).

Figure 3 with 1 supplement see all

Download asset Open asset

We found that 20–40 Hz power significantly increased with performance level (LMEM ANOVA significant effect of session: F_2,4812 = 46.71, p<0.001), which complements our previous study showing learning-related differences in waveform amplitude (Allen et al., 2016). When collapsing across electrodes, the well-trained session showed significantly higher power compared to novel 1 (t₄₈₁₂=8.451, p<0.001), while power during the novel 1 and novel 2 sessions did not significantly differ (t₄₈₁₂=1.470, p=0.14168). Notably, this performance effect did not significantly vary across the proximodistal axis (nonsignificant interaction between electrode and session F_6,4806 = 0.745, p=0.614), but instead scaled with the local amplitude of 20–40 Hz, suggesting that this behavioral correlate is present throughout dorsal CA1. With regard to the main effect of electrode (collapsing across sessions), we observed a significant difference across sites (LMEM ANOVA: F_3,4812 = 52.784, p<0.001). More precisely, average power was significantly higher in electrode 1 compared to electrodes 2–4 (t₄₈₁₂ = –2.969, p=0.003, t₄₈₁₂ = –8.896, p<0.001, and t₄₈₁₂ = –11.095, p<0.001, respectively) (Figure 3C). It is important to note that we chose an a priori time window (−250–0 ms prior to port withdrawal) to quantify this performance effect. We also used a data-driven approach to identify epochs during which 20–40 Hz dynamics statistically vary with performance level (Figure 3B), which showed that the performance effect was not limited to the window defined a priori.

Lastly, we tested whether the observed performance-related effects are specific to 20–40 Hz dynamics or generalize to other bands, including theta (4–12 Hz), the slow gamma range nonoverlapping with 20–40 Hz (40–55 Hz), and fast gamma (65–100 Hz; Figure 3—figure supplement 1). We found that performance-related effects were weaker, and the patterns differed from those of the 20–40 Hz range. Whereas 20–40 Hz power increased with performance, theta and fast gamma power decreased (theta LMEM ANOVA: F_2,4806 = 7.681, p<0.001; fast gamma LMEM ANOVA: F_2,4806 = 8.147, p<0.001) and slow gamma showed a nonlinear relationship with session performance (LMEM ANOVA: F_2,4815 = 5.993, p=0.003). This suggests that the 20–40 Hz band is unique in showing a distinctive increase in power with performance level.

20–40 Hz steady-state power in proximal CA1 is predictive of accuracy for temporal order judgments

To shed light on the type of processing reflected by 20–40 Hz oscillations, we took advantage of the four different trial types included in our paradigm: InSeq trials that were correctly or incorrectly identified (InSeq+, InSeq-) and OutSeq trials correctly or incorrectly identified (OutSeq+, OutSeq-) (Figure 1C). We began by visualizing power dynamics as a function of the entire time course of trials, focusing on the electrode and session with highest 20–40 Hz power (electrode 1, well-trained session). Aligning data to port entry (Figure 4A, left), we observed that the rising phase (~0–750 ms) was similar across trial types and that differences emerged later in the trial. Specifically, we observed that power was sustained (reaching steady-state) on trials in which animals performed a ‘hold’response (InSeq+ and OutSeq-), whereas it fell off earlier (during the transient phase) on trials in which the animals performed a ‘withdraw’ response (OutSeq+ and InSeq-). Aligning data to port withdrawal confirmed that the withdrawal response occurs during steady-state for the hold responses, but stochastically during the transient response or early steady-state for the withdraw responses (Figure 4A, right).

Figure 4 with 1 supplement see all

Download asset Open asset

To account for the confounding effect of this post hoc observation, we first compared power across the four trial types using an early epoch relative to port entry (250–500 ms). This epoch was chosen with the aim of minimizing the possibility of capturing modulations in the power dynamics by hold-/withdraw-response behavior. There were no significant differences between trial types in this epoch (LMEM ANOVA nonsignificant effect of trial type: F_3,749 = 0.480, p=0.6935; Figure 4C, left), suggesting that the 20–40 Hz transient phase did not differ with response type (hold/withdraw) nor accuracy (correct/incorrect). Although there was a significant effect of trial type when focusing on power during the 250 ms period preceding port withdrawal (LMEM ANOVA: F_3,743 = 9.451, p<0.001; Figure 4C, right), this effect appears primarily due to the post hoc observation mentioned above (as well as the variability observed on InSeq- trials). In fact, InSeq+ trials (hold responses) were higher than OutSeq+ and InSeq- (withdraw responses) trial types (t₇₄₃ = –2.829, p=0.005 and t₇₄₃ = –4.787, p<0.001, respectively), but power in this epoch did not significantly differ between the hold-response trials (InSeq+ and OutSeq-) (t₇₄₃ = –0.458, p=0.647).

To more thoroughly examine whether 20–40 Hz power is associated with accurate temporal order judgments, we used a random forest classifier to test whether power in a set of predefined epochs and across the CA1 axis predicts response accuracy. Given our post hoc observation of fundamentally different dynamics for the hold- vs. withdraw-response trials, all classification analyses focused on comparing accuracy using the hold-response trial types (InSeq+ vs. OutSeq). This enabled comparisons between correct and incorrect trials while matching motor behavior and nosepoke duration distributions. Classification analysis was first performed using a 16-D feature vector that included 20–40 Hz power during two epochs relative to port entry (200–350, 350–500 ms) and two epochs relative to port withdrawal (−400 to –200, –200 to 0 ms) in each of the four CA1 sites (4 epochs × 4 electrodes) (Figure 4—figure supplement 1A, left). We found that these features were not sufficient to discriminate correct and incorrect trials as classifier performance was at chance levels (mean area under curve [AUC], true positive rate [TPR], and false positive rate [FPR] of 0.492, 0.497, and 0.474, respectively) (Figure 4—figure supplement 1A, right). As a follow-up, we then used a larger feature space that included time epochs covering the majority of the waveform. We defined a 24-D feature vector that included 20–40 Hz power in four epochs relative to port entry (200–350, 350–500, 500–650 and 650–1200 ms; the latter two not included in the previous analysis) and two epochs relative to port withdrawal (−400 to –200, –200 to 0 ms) in each of the four CA1 sites (6 time epochs × 4 electrodes) (Figure 4—figure supplement 1B, left). Using the overlap of lasso (alpha = 0, 0.01, 0.02, and 0.03) and Gini impurity-based relative importance (Figure 4—figure supplement 1C, D) , the 24-D vector was reduced to three features most predictive of trial accuracy: 20–40 Hz power during the 650–1200 ms period relative to port entry, in the most proximal sites (electrodes 1–3) (Figure 4—figure supplement 1E).

We found that these three features were sufficient to discriminate between correct and incorrect temporal order judgments (InSeq+ and OutSeq-, respectively). In fact, the classifier performed with mean areas under the curve (AUC) of 0.588 and 0.477 on the observed and shuffled data, respectively (Figure 4D), and a Mann–Whitney U test showed a statistically significant difference between the observed and null AUC values (with higher classification performance on the observed data; U = 839,538, p<0.001). Observed compared to null classification receiver operating characteristic (ROC) curve is displayed in Figure 4D, with mean observed classifier AUC, TPR, and FPR of 0.588, 0.562, and 0.444, respectively. It is important to note that this prediction analysis uses pooled data across animals.

Post hoc group analyses of the identified feature (650–1200 ms epoch) showed a significant effect of trial type (InSeq+ vs. OutSeq-; LMEM ANOVA: F_1,2406 = 32.687, p<0.001) and electrode (F_3,2406 = 90.573, p<0.001), but no significant interaction (F_3,2406 = 0.193, p=0.901) (Figure 4E). Interestingly, power was significantly higher on incorrect temporal order judgments (OutSeq- power was higher than InSeq+). Across trial types, power was significantly higher in electrode 1 compared to electrodes 3 and 4 (t₂₄₀₆ = –12.310, p<0.001 and t₂₄₀₆ = –13.025, p<0.001, respectively). Power in electrode 1 was also higher than power in electrode 2; however, this difference was not significant (t₂₄₀₆ = –2.215, p=0.027). Altogether, these results demonstrate that 20–40 Hz steady-state power in proximal CA1 is predictive of accuracy for temporal order judgments during hold-response trials, and that incorrect judgments are associated with significantly higher 20–40 Hz power.

The increase in CA1 20–40 Hz power with session performance is selective to hold-response trials

Figure 3 shows increased 20–40 Hz power with session performance, and Figure 4 shows that 20–40 Hz dynamics that predict trial accuracy occur late in the trial. In linking these results, we aimed to delineate between two possibilities. The first possibility is that increased power with session performance reflects a direct index of sequence knowledge (i.e., CA1 representation of sequence knowledge or retrieval thereof). This would be supported if increased power with performance were selective to correct trial types (InSeq+ and OutSeq+). The second possibility is that, given that accuracy predictive dynamics occurred later in the trial (after memory for order has been retrieved and a post-retrieval decision made), power increases reflect a post-decision state that varies with session performance. This would be supported if the session performance effect were selective to trials reaching a post-decision state (hold responses, InSeq+, OutSeq-).

To test these two possibilities, we investigated differences in dynamics across sessions separately for the four trial types (Figure 5). Consistent with the latter prediction, our cluster-based permutation testing (CBPT) identified epochs when 20–40 Hz power significantly increased with performance level (excluding data from novel 1 session for sampling reasons). This performance effect was observed for the hold-response trial types (InSeq+, OutSeq-; Figure 5A) but not the withdraw-response trial types (Figure 5B; which included OutSeq+ trials where order memory retrieval is presumably intact and improved with training). To directly compare power in a fixed epoch across all trial types, the OutSeq- epoch was used since it represented the overlap in significant epochs between the two hold-response trial types. During this epoch, a significant effect of performance level was observed for both hold-response trial types (InSeq+ and OutSeq-; LMEM ANOVA significant effect of performance: F_2,1229 = 14.184, p<0.001 and F_2,173 = 12.072, p<0.001, respectively). For both trial types, epoch power during the well-trained session was significantly higher than that of the novel 1 session (t₁₂₂₉ = 3.984, p<0.001 and t₁₇₃ = 4.462, p<0.001, respectively). Lastly, for both trial types, no significant differences in epoch power between novel 1 and novel 2 sessions were observed (t₁₂₂₉ = –0.438, p=0.662 and t₁₇₃ = 0.748, p=0.455). In contrast, epoch power during both withdraw-response trial types (OutSeq+ and InSeq-) did not significantly differ with performance (LMEM ANOVA nonsignificant effect of performance: F_2,128 = 0.384, p=0.682 and F_2,116 = 0.796, p=0.453, respectively). These results support the notion that power increases with session performance may reflect a post-decision state that varies with knowledge of the sequence.

Figure 5

Download asset Open asset

We collected LFP activity from the CA1 hippocampal region as rats performed a sequence memory task to identify the frequency content supporting nonspatial information processing. The data presented here expand upon our previous report of 20–40 Hz power during the odor sequence processing periods of the task (Allen et al., 2016) by evaluating the behavioral relevance and anatomical distribution of this rhythm across the CA1 proximodistal axis, as well as providing a direct comparison with another behavioral state (running). First, we demonstrated that running and odor sequence processing epochs were characterized by different spectral features. Running was associated with increased power in theta (4–12 Hz) and >24 Hz frequency ranges, whereas odor sequence processing was associated with increased power in theta (at a lower center frequency than during running) and 20–40 Hz ranges. Second, we showed that in both behavioral states there were distinct gradients of recruited frequency power along the CA1 proximodistal axis. During odor processing, 20–40 Hz power was higher in proximal CA1, whereas theta was higher in distal CA1. During running periods, theta was higher in proximal, whereas >24 Hz power was higher in distal CA1. Third, we observed differential performance-related effects across frequency bands; 20–40 Hz power dynamics were positively associated with performance across sessions, while theta and fast gamma demonstrated a negative association. Fourth, we observed differential power dynamics associated with the task’s discriminative response requirements. Both hold-response trials (InSeq+ and OutSeq-) and withdraw-response trials (OutSeq+ and InSeq-) displayed a transient phase where 20–40 Hz power gradually increased at a similar rate. However, dynamics subsequently diverged; 20–40 Hz power declined on withdraw-response (presumed OutSeq decisions) trials while hold-response trials (presumed InSeq decisions) displayed a steady-state period of elevated 20–40 Hz power that remained until the animal withdrew. Fifth, we identified the steady phase of 20–40 Hz power in proximal CA1 as a key component carrying task-critical information. In fact, this component of the signal (observed on hold-response trials) increased with session performance and was predictive of the accuracy of temporal order judgments. Collectively, these results suggest that 20–40 Hz power dynamics in the proximal segment of CA1 are linked to processing the temporal order of nonspatial events and reflect a post-decision state that may serve to maintain and protect trial-specific information until a response can be performed.

The nature of this experimental paradigm led to two potential limitations to consider when interpreting the findings. First, the use of a nonspatial discriminative response (hold/withdraw) prevented us from directly equating response duration across all trial types and unexpectedly led to different power dynamics between hold- and withdraw-response trials (possibly attributable to differences in cue utilization across conditions, as in Leventhal et al., 2012, or merely in trial duration). For that reason, we focused our secondary analyses on the effects of response accuracy to hold-response trials. Second, to ensure adequate performance, the task requires that the number of OutSeq trials be kept relatively low (otherwise it is unclear which sequence is being tested). This results in an uneven number of observations across trial types that could disproportionally influence a subset of our analyses. However, we controlled for this possibility by conducting pooled analyses using an LMEM ANOVA to account for correlations in measures obtained from the same subject and by matching sample sizes in the classification analyses. Overall, we believe that these approaches significantly mitigated the potential influence of these confounding factors on the interpretation of our results.

Our results expand our understanding of CA1 spectral profiles across different behavioral states and conditions. First, prior reports have shown that theta and broadband gamma power are recruited during spatial exploration and that power dynamics vary with behavior (Trimper et al., 2017; Chen et al., 2011; Ahmed and Mehta, 2012). We expand on these observations by demonstrating gradients of theta and gamma power along the CA1 proximodistal axis during running and by revealing that a different spectral profile is observed during nonspatial behavior. Second, previous studies have shown a similar recruitment of 20–40 Hz oscillations in odor-based tasks (Martin et al., 2007; Igarashi et al., 2014; Rangel et al., 2016). Oscillations in a similar range (15–35 Hz) were recorded in the hippocampus and olfactory bulb of rats performing a go/no-go odor discrimination task (Martin et al., 2007). Interestingly, that study showed a learning-related increase in oscillatory power in the olfactory bulb as well as in the coherence between signals from bulb and hippocampus. However, oscillatory power in the hippocampus did not change as a function of learning. Similarly, 20–40 Hz oscillations were observed in distal CA1 (dCA1) and lateral entorhinal cortex (LEC) of rats engaged in an odor-place association task (Igarashi et al., 2014). In that study, learning-related increases in dCA1-LEC coherence were observed, but oscillatory power in either region did not significantly change with learning. This suggests that although these oscillations are observed in a variety of odor-based tasks, increases in odor familiarity or task performance do not necessarily result in a predictable increase in power. Instead, our findings suggest that 20–40 Hz dynamics are linked to specific task demands and, in this case, reflect a preferential engagement of proximal CA1 in processing temporal order judgments.

Evidence for functional heterogeneity along the CA1 proximodistal axis has been previously reported (Henriksen et al., 2010; Hartzell et al., 2013; Nakazawa et al., 2016; Ng et al., 2018). Such heterogeneity is not surprising given the known connectivity gradients of the lateral and medial entorhinal cortices (LEC and MEC; van Strien et al., 2009; Witter et al., 2017), namely, that proximal and distal CA1 are more anatomically and functionally associated with MEC and LEC, respectively. Consequently, the observation of different proximodistal gradients of oscillatory power across studies may result from differences in task demands, which could promote differential engagement of these entorhinal-CA1 circuits (although other factors, including methodological and analytical approaches, may also contribute to this difference). For instance, Igarashi et al., 2014 showed that 20–40 Hz power and LEC-CA1 coherence were higher in distal CA1 during performance of an odor-place association task (whereas here we show higher power in proximal CA1). Since performance in their task depends on the correct identification of the specific perceptual features that distinguish one odor from another, the power increase in distal CA1 may reflect a stronger engagement of the LEC-dCA1 component of the circuit (as LEC receives strong olfactory input, including direct projections from the olfactory bulb; Haberly and Price, 1978; Agster and Burwell, 2009). In contrast, in our experiment, identification of the presented odor alone is insufficient for correct performance – the animal must further identify whether the odor is being presented in the appropriate temporal context. The power increase in proximal CA1 we observed may reflect the fact that this additional temporal requirement preferentially engaged the MEC-pCA1 component of the circuit. Interestingly, although MEC is typically associated with spatial navigation functions, this interpretation is supported by a recent study by Robinson et al., 2017, in which they demonstrated that optogenetic inactivation of MEC disrupted temporal coding in CA1, while sparing spatial and object coding. Together with the proximodistal pattern we observed, these results suggest that perhaps the MEC→proximal CA1 microcircuit may be important for the processing of nonspatial temporal information.

Oscillations in the 20–40 Hz range (typically referred to as beta) occur widely across the brain and have been associated with different functions across brain regions (Engel and Fries, 2010; Schmidt et al., 2019). Of particular interest here is the hypothesis that beta serves a gating function that helps lock in and protect a neural state representing trial-critical information (i.e., maintain the status quo; Engel and Fries, 2010; Weiss and Mueller, 2012). Such a mechanism would be disruptive during encoding or recall processes but advantageous for maintaining a decision until the corresponding response is made, which is consistent with the 20–40 Hz trial-level dynamics we observed in CA1 (low power early in the odor period, but high power in the late period before the response). More precisely, as OutSeq+ trials required animals to perform an ‘immediate withdraw-response’ (i.e., not wait until the signal), we can use the timing of those responses to infer when temporal order decisions occurred. During those trials, it was the transient phase of 20–40 Hz power that was consistently observed; any encoding and/or recall, as well as the decision based on such processes, necessarily occurred prior to or early in the induction of 20–40 Hz power. This is in line with observations that beta power is attenuated during encoding (Lundqvist et al., 2016; Zavala et al., 2017) and complements findings that transcranial magnetic stimulation at the beta frequency impairs encoding (Hanslmayr et al., 2014). The perspective that beta helps maintain the status quo would predict that the steady-state 20–40 Hz period observed on hold-response trials reflects the system actively maintaining an ensemble or network state in support of the decision to perform a hold response. Though it is possible the post-decision ensemble state reflects a general signal, such as a postural one or a nonspecific signal of a temporal context match, our current finding showed that this epoch is predictive of accuracy and our previous work showed that specific odor and temporal context information is represented at the single-cell (Allen et al., 2016) and ensemble (Shahbaba et al., 2022) levels during this late trial period. It is therefore possible that what is ‘locked in’ during the steady-state 20–40 Hz period is the trial-specific content supporting the animal’s judgment that the odor is presented in the correct temporal context.

This potential gating function of beta is also consistent with the performance-related effects we observed at the session level. Here, we showed that 20–40 Hz power increases with session performance (lowest in novel 1, highest in well-trained). High beta during the well-trained session possibly reflects higher levels of gating and effective suppression of distracting processes; this is a high-performing session in which the sequence is already well-learned, minimizing the need for additional knowledge acquisition. In contrast, lower beta power in the novel 1 and 2 sessions, which are associated with a higher demand for encoding, may enable flexible knowledge acquisition processes that would otherwise have been interrupted. Consistent with complementary dynamics of beta and gamma rhythms previously reported (Howe et al., 2011; Lundqvist et al., 2016), we find an inverse pattern for fast gamma power (and theta) with performance, whereby fast gamma power is higher during acquisition of a new sequence (novel 1–2) and decreases with session performance. This finding highlights the distinction between the functional role of 20–40 Hz and that of the other rhythms.

Interestingly, when considering the relationship between steady-state 20–40 Hz power and performance at both the trial and session levels, we noted a pattern more complex than a simple linear relationship. Whereas 20–40 Hz power increased with session performance (from novel 1 to well-trained), steady-state power in the well-trained session was higher during OutSeq- (incorrect hold response) compared to InSeq+ (correct hold response). Although this result may seem counterintuitive, this effect is compatible with the proposed gating function of beta. Specifically, that beta serves to coordinate integration of exogenous information with endogenous information (Iversen et al., 2009; Bastiaansen et al., 2010), and that beta power regulates the degree of suppression of seemingly distracting processes (Lundqvist et al., 2016; Schmidt et al., 2019; Miller et al., 2018). Proximal CA1 higher steady-state beta levels during incorrect trials may reflect over-suppression, leading to interference with processes needed for successful performance on the task. In other words, the appropriate level of suppression may be advantageous in attenuating distractions, but too high a level may impede essential processes. Consistent with this are clinical reports demonstrating that increased beta in patients with Parkinson’s disease is associated with highly sustained, rigid movements (Chen et al., 2007; Pogosyan et al., 2009; Brown, 2007) and difficulty in switching behavioral paradigm (Stoffers et al., 2001); clinically disrupting beta power improves these symptoms, with a side effect of inability to ignore distractions and difficulty in focusing on the task at hand (Cools et al., 2003; Moustafa et al., 2008). Together, these findings point to graded effects of beta on performance extending to CA1. Alternatively, beta effects may reflect more of a threshold process: higher steady-state power may represent a point of no return, despite retrieval of knowledge that suggests a change in behavior. This latter possibility is consistent with the observation made by Leventhal et al., 2012, in which animals were given a second stop cue following an initial go cue. The likelihood of stopping decreased when the second cue occurred during the beta synchronization event elicited from the first cue, reflecting a ‘closed gate’ on the behavioral state initially cued. Future studies are needed to adjudicate between these possibilities, possibly by specifying the relationship between the information content represented in the ensemble activity and the 20–40 Hz power dynamics.

Future studies will also be needed to determine the degree to which our findings relate to cue utilization. Utilization of a sensory cue to inform motor output has been shown to recruit ~20 Hz (beta) in cortico-basal ganglia circuits (CBG, Leventhal et al., 2012), CA1-striatal circuits (Lansink et al., 2016), and the dentate gyrus (DG; Rangel et al., 2015), giving rise to a working hypothesis for the functional role of ~20 Hz in sensorimotor integration. Since, in this study, animals utilized the odor cue to either maintain or terminate a hold posture, it is possible that this aspect of sensorimotor integration contributed to the oscillations we observed. Unfortunately, resolving this issue would require a different experimental design. Similarly, further work will be needed to determine the origin of the 20–40 Hz rhythm in CA1. CA1 receives input from a number of other sources, including entorhinal cortex, CA2, CA3, and the medial septum (van Strien et al., 2009), and the contributions of these upstream structures to the generation of this oscillation in CA1 remain unknown. Though we previously showed that the spiking activity of a proportion of CA1 neurons is phase-locked to these 20–40 Hz oscillations (Allen et al., 2016), we cannot rule out the possibility that volume conduction contributed to the presence of these rhythms and their gradients. A prior study performed current source density analysis for hippocampal 15–35 Hz power, spanning DG, CA3, and CA1, and localized this rhythm to the DG (Rangel et al., 2015). Altogether, future experiments are needed to clarify the relative contribution of mnemonic and sensorimotor processes to the observed rhythms and their gradients, as well as the potential influence of volume conduction.

Finally, evidence from multisite recordings suggest the 20–40 Hz dynamics we observed in CA1 reflect task-critical, multiregional integration. As mentioned earlier, coherence in the beta range between the hippocampus and olfactory regions (Martin and Ravel, 2014; Kay, 2014), as well as with the entorhinal cortex (Igarashi et al., 2014), has been previously reported during odor-dependent tasks. Importantly, these dynamics are not restricted to olfactory processing (visual cues: Lansink et al., 2016; auditory cues: Leventhal et al., 2012) nor do they exclusively coordinate hippocampus with olfactory or medial temporal lobe regions. Lansink et al., 2016 demonstrated CA1 beta (15–20 Hz) recruitment as rats responded in the presence of a reward-predictive visual cue that directed navigation (as opposed to goal-directed navigation absent a cue). Notably, during this recruitment neuronal firing was synchronized between CA1 and ventral striatum (Lansink et al., 2016). Specific to our task, structures connected to the hippocampus have also been shown to be critical for performance, including prefrontal, perirhinal, and mid-thalamic regions (Fortin et al., 2016; Jayachandran et al., 2019). As each structure likely contains region-specific representations, beta may serve to bind these regions together to coordinate these distinct task-critical representations across regions and support goal-directed behavior (Fries, 2005). To clarify, multiregional integration may serve as the general function of beta, which, in our paradigm, is in the service of supporting order memory judgments. Increases in CA1 beta have been reported in other task conditions and demands, such as exploration and novelty (Berke et al., 2008; França et al., 2014), which could reflect multiregional integration supporting other forms of task-critical information processing. Collectively, these findings provide evidence for the far-reaching role of beta in coordinating information processing across brain regions by gating their region-specific patterns of activity important for the task at hand.

Our group previously published using the same dataset, and a detailed description of the methods can be found in Allen et al., 2016. The methods are summarized below.

Data available on Dryad, Data DOI: https://doi.org/10.7280/D11960.

1. 1. Gattas S
   2. Elias G
   3. Janecek J
   4. Yassa MA
   5. Fortin NJ

   (2022) Dryad Digital Repository

   Proximal CA1 20-40 Hz power dynamics reflect trial-specific information processing supporting nonspatial sequence memory.

   https://doi.org/10.7280/dryad.D11960

1. 1. Agster KL
   2. Burwell RD

   (2009) Cortical efferents of the perirhinal, postrhinal, and entorhinal cortices of the rat

   Hippocampus 19:1159–1186.

   https://doi.org/10.1002/hipo.20578

   • PubMed
   • Google Scholar
2. 1. Ahmed OJ
   2. Mehta MR

   (2012) Running speed alters the frequency of hippocampal gamma oscillations

   The Journal of Neuroscience 32:7373–7383.

   https://doi.org/10.1523/JNEUROSCI.5110-11.2012

   • PubMed
   • Google Scholar
3. 1. Allen TA
   2. Fortin NJ

   (2013) The evolution of episodic memory

   PNAS 110 Suppl 2:10379–10386.

   https://doi.org/10.1073/pnas.1301199110

   • PubMed
   • Google Scholar
4. 1. Allen TA
   2. Morris AM
   3. Mattfeld AT
   4. Stark CEL
   5. Fortin NJ

   (2014) A Sequence of events model of episodic memory shows parallels in rats and humans

   Hippocampus 24:1178–1188.

   https://doi.org/10.1002/hipo.22301

   • PubMed
   • Google Scholar
5. 1. Allen TA
   2. Salz DM
   3. McKenzie S
   4. Fortin NJ

   (2016) Nonspatial Sequence Coding in CA1 Neurons

   The Journal of Neuroscience 36:1547–1563.

   https://doi.org/10.1523/JNEUROSCI.2874-15.2016

   • PubMed
   • Google Scholar
6. 1. Bastiaansen M
   2. Magyari L
   3. Hagoort P

   (2010) Syntactic unification operations are reflected in oscillatory dynamics during on-line sentence comprehension

   Journal of Cognitive Neuroscience 22:1333–1347.

   https://doi.org/10.1162/jocn.2009.21283

   • PubMed
   • Google Scholar
7. 1. Berke JD
   2. Hetrick V
   3. Breck J
   4. Greene RW

   (2008) Transient 23-30 Hz oscillations in mouse hippocampus during exploration of novel environments

   Hippocampus 18:519–529.

   https://doi.org/10.1002/hipo.20435

   • PubMed
   • Google Scholar
8. 1. Brown P

   (2007) Abnormal oscillatory synchronisation in the motor system leads to impaired movement

   Current Opinion in Neurobiology 17:656–664.

   https://doi.org/10.1016/j.conb.2007.12.001

   • PubMed
   • Google Scholar
9. 1. Buzsáki G
   2. Logothetis N
   3. Singer W

   (2013) Scaling brain size, keeping timing: evolutionary preservation of brain rhythms

   Neuron 80:751–764.

   https://doi.org/10.1016/j.neuron.2013.10.002

   • PubMed
   • Google Scholar
10. 1. Buzsáki G.
    2. Moser EI

    (2013) Memory, navigation and theta rhythm in the hippocampal-entorhinal system

    Nature Neuroscience 16:130–138.

    https://doi.org/10.1038/nn.3304

    • PubMed
    • Google Scholar
11. 1. Buzsáki G.
    2. Tingley D

    (2018) Space and Time: The Hippocampus as a Sequence Generator

    Trends in Cognitive Sciences 22:853–869.

    https://doi.org/10.1016/j.tics.2018.07.006

    • PubMed
    • Google Scholar
12. 1. Chen CC
    2. Litvak V
    3. Gilbertson T
    4. Kühn A
    5. Lu CS
    6. Lee ST
    7. Tsai CH
    8. Tisch S
    9. Limousin P
    10. Hariz M
    11. Brown P

    (2007) Excessive synchronization of basal ganglia neurons at 20 Hz slows movement in Parkinson’s disease

    Experimental Neurology 205:214–221.

    https://doi.org/10.1016/j.expneurol.2007.01.027

    • PubMed
    • Google Scholar
13. 1. Chen Z
    2. Resnik E
    3. McFarland JM
    4. Sakmann B
    5. Mehta MR
    6. Borst A

    (2011) Speed Controls the Amplitude and Timing of the Hippocampal Gamma Rhythm

    PLOS ONE 6:e21408.

    https://doi.org/10.1371/journal.pone.0021408

    • PubMed
    • Google Scholar
14. 1. Colgin LL
    2. Denninger T
    3. Fyhn M
    4. Hafting T
    5. Bonnevie T
    6. Jensen O
    7. Moser MB
    8. Moser EI

    (2009) Frequency of gamma oscillations routes flow of information in the hippocampus

    Nature 462:353–357.

    https://doi.org/10.1038/nature08573

    • PubMed
    • Google Scholar
15. 1. Colgin LL

    (2016) Rhythms of the hippocampal network

    Nature Reviews. Neuroscience 17:239–249.

    https://doi.org/10.1038/nrn.2016.21

    • PubMed
    • Google Scholar
16. 1. Cools R
    2. Barker RA
    3. Sahakian BJ
    4. Robbins TW

    (2003) L-Dopa medication remediates cognitive inflexibility, but increases impulsivity in patients with Parkinson’s disease

    Neuropsychologia 41:1431–1441.

    https://doi.org/10.1016/S0028-3932(03)00117-9

    • PubMed
    • Google Scholar
17. 1. Drieu C
    2. Zugaro M

    (2019) Hippocampal Sequences During Exploration: Mechanisms and Functions

    Frontiers in Cellular Neuroscience 13:232.

    https://doi.org/10.3389/fncel.2019.00232

    • PubMed
    • Google Scholar
18. 1. Eichenbaum H

    (2014) Time cells in the hippocampus: A new dimension for mapping memories

    Nature Reviews. Neuroscience 15:732–744.

    https://doi.org/10.1038/nrn3827

    • PubMed
    • Google Scholar
19. 1. Engel AK
    2. Fries P

    (2010) Beta-band oscillations--signalling the status quo?

    Current Opinion in Neurobiology 20:156–165.

    https://doi.org/10.1016/j.conb.2010.02.015

    • PubMed
    • Google Scholar
20. 1. Fortin NJ
    2. Agster KL
    3. Eichenbaum HB

    (2002) Critical role of the hippocampus in memory for sequences of events

    Nature Neuroscience 5:458–462.

    https://doi.org/10.1038/nn834

    • PubMed
    • Google Scholar
21. 1. Fortin NJ
    2. Asem JSA
    3. Ng CW
    4. Quirk CR
    5. Allen TA
    6. Elias GA

    (2016) Distinct contributions of hippocampal, prefrontal, perirhinal and nucleus reuniens regions to the memory for sequences of events

    Society for Neuroscience Abstracts (San Diego, CA 1:2215.

    https://doi.org/10.1016/j.bbr.2019.112215

    • Google Scholar
22. 1. Foster DJ
    2. Knierim JJ

    (2012) Sequence learning and the role of the hippocampus in rodent navigation

    Current Opinion in Neurobiology 22:294–300.

    https://doi.org/10.1016/j.conb.2011.12.005

    • PubMed
    • Google Scholar
23. 1. França ASC
    2. do Nascimento GC
    3. Lopes-dos-Santos V
    4. Muratori L
    5. Ribeiro S
    6. Lobão-Soares B
    7. Tort ABL

    (2014) Beta2 oscillations (23-30 Hz) in the mouse hippocampus during novel object recognition

    The European Journal of Neuroscience 40:3693–3703.

    https://doi.org/10.1111/ejn.12739

    • Google Scholar
24. 1. Fries P

    (2005) A mechanism for cognitive dynamics: neuronal communication through neuronal coherence

    Trends in Cognitive Sciences 9:474–480.

    https://doi.org/10.1016/j.tics.2005.08.011

    • Google Scholar
25. 1. Haberly LB
    2. Price JL

    (1978) Association and commissural fiber systems of the olfactory cortex of the rat. I. Systems originating in the piriform cortex and adjacent areas

    The Journal of Comparative Neurology 178:711–740.

    https://doi.org/10.1002/cne.901780408

    • PubMed
    • Google Scholar
26. 1. Hanslmayr S
    2. Matuschek J
    3. Fellner MC

    (2014) Entrainment of Prefrontal Beta Oscillations Induces an Endogenous Echo and Impairs Memory Formation

    Current Biology 24:904–909.

    https://doi.org/10.1016/j.cub.2014.03.007

    • PubMed
    • Google Scholar
27. 1. Hartzell AL
    2. Burke SN
    3. Hoang LT
    4. Lister JP
    5. Rodriguez CN
    6. Barnes CA

    (2013) Transcription of the Immediate-Early Gene Arc in CA1 of the Hippocampus Reveals Activity Differences along the Proximodistal Axis That Are Attenuated by Advanced Age

    Journal of Neuroscience 33:3424–3433.

    https://doi.org/10.1523/JNEUROSCI.4727-12.2013

    • PubMed
    • Google Scholar
28. 1. Henriksen EJ
    2. Colgin LL
    3. Barnes CA
    4. Witter MP
    5. Moser MB
    6. Moser EI

    (2010) Spatial Representation along the Proximodistal Axis of CA1

    Neuron 68:127–137.

    https://doi.org/10.1016/j.neuron.2010.08.042

    • PubMed
    • Google Scholar
29. 1. Howe MW
    2. Atallah HE
    3. McCool A
    4. Gibson DJ
    5. Graybiel AM

    (2011) Habit learning is associated with major shifts in frequencies of oscillatory activity and synchronized spike firing in striatum

    PNAS 108:16801–16806.

    https://doi.org/10.1073/pnas.1113158108

    • PubMed
    • Google Scholar
30. 1. Igarashi KM
    2. Lu L
    3. Colgin LL
    4. Moser MB
    5. Moser EI

    (2014) Coordination of entorhinal–hippocampal ensemble activity during associative learning

    Nature 510:143–147.

    https://doi.org/10.1038/nature13162

    • PubMed
    • Google Scholar
31. 1. Iversen JR
    2. Repp BH
    3. Patel AD

    (2009) Top-Down Control of Rhythm Perception Modulates Early Auditory Responses

    Annals of the New York Academy of Sciences 1169:58–73.

    https://doi.org/10.1111/j.1749-6632.2009.04579.x

    • PubMed
    • Google Scholar
32. 1. Jayachandran M
    2. Linley SB
    3. Schlecht M
    4. Mahler SV
    5. Vertes RP
    6. Allen TA

    (2019) Prefrontal Pathways Provide Top-Down Control of Memory for Sequences of Events

    Cell Reports 28:640–654.

    https://doi.org/10.1016/j.celrep.2019.06.053

    • Google Scholar
33. 1. Kay LM

    (2014) Circuit oscillations in odor perception and memory

    Progress in Brain Research 208:223–251.

    https://doi.org/10.1016/B978-0-444-63350-7.00009-7

    • PubMed
    • Google Scholar
34. 1. Kesner RP
    2. Gilbert PE
    3. Barua LA

    (2002) The role of the hippocampus in memory for the temporal order of a sequence of odors

    Behavioral Neuroscience 116:286–290.

    https://doi.org/10.1037//0735-7044.116.2.286

    • PubMed
    • Google Scholar
35. 1. Lansink CS
    2. Meijer GT
    3. Lankelma JV
    4. Vinck MA
    5. Jackson JC
    6. Pennartz CMA

    (2016) Reward Expectancy Strengthens CA1 Theta and Beta Band Synchronization and Hippocampal-Ventral Striatal Coupling

    The Journal of Neuroscience 36:10598–10610.

    https://doi.org/10.1523/JNEUROSCI.0682-16.2016

    • PubMed
    • Google Scholar
36. 1. Leventhal DK
    2. Gage GJ
    3. Schmidt R
    4. Pettibone JR
    5. Case AC
    6. Berke JD

    (2012) Basal ganglia beta oscillations accompany cue utilization

    Neuron 73:523–536.

    https://doi.org/10.1016/j.neuron.2011.11.032

    • PubMed
    • Google Scholar
37. 1. Lisman JE

    (1999) Relating hippocampal circuitry to function: recall of memory sequences by reciprocal dentate-CA3 interactions

    Neuron 22:233–242.

    https://doi.org/10.1016/s0896-6273(00)81085-5

    • PubMed
    • Google Scholar
38. 1. Lundqvist M
    2. Rose J
    3. Herman P
    4. Brincat SLL
    5. Buschman TJJ
    6. Miller EKK

    (2016) Gamma and Beta Bursts Underlie Working Memory

    Neuron 90:152–164.

    https://doi.org/10.1016/j.neuron.2016.02.028

    • PubMed
    • Google Scholar
39. 1. MacDonald CJ
    2. Lepage KQ
    3. Eden UT
    4. Eichenbaum H

    (2011) Hippocampal “time cells” bridge the gap in memory for discontiguous events

    Neuron 71:737–749.

    https://doi.org/10.1016/j.neuron.2011.07.012

    • PubMed
    • Google Scholar
40. 1. Martin C
    2. Beshel J
    3. Kay LM

    (2007) An olfacto-hippocampal network is dynamically involved in odor-discrimination learning

    Journal of Neurophysiology 98:2196–2205.

    https://doi.org/10.1152/jn.00524.2007

    • PubMed
    • Google Scholar
41. 1. Martin C
    2. Ravel N

    (2014) Beta and gamma oscillatory activities associated with olfactory memory tasks: different rhythms for different functional networks?

    Frontiers in Behavioral Neuroscience 8:218.

    https://doi.org/10.3389/fnbeh.2014.00218

    • Google Scholar
42. 1. McNaughton BL
    2. Morris RGM

    (1987) Hippocampal synaptic enhancement and information storage within a distributed memory system

    Trends in Neurosciences 10:408–415.

    https://doi.org/10.1016/0166-2236(87)90011-7

    • Google Scholar
43. 1. Miller EK
    2. Lundqvist M
    3. Bastos AM

    (2018) Working Memory 2.0

    Neuron 100:463–475.

    https://doi.org/10.1016/j.neuron.2018.09.023

    • PubMed
    • Google Scholar
44. 1. Moustafa AA
    2. Sherman SJ
    3. Frank MJ

    (2008) A dopaminergic basis for working memory, learning and attentional shifting in Parkinsonism

    Neuropsychologia 46:3144–3156.

    https://doi.org/10.1016/j.neuropsychologia.2008.07.011

    • PubMed
    • Google Scholar
45. 1. Nakazawa Y
    2. Pevzner A
    3. Tanaka KZ
    4. Wiltgen BJ

    (2016) Memory retrieval along the proximodistal axis of CA1

    Hippocampus 26:1140–1148.

    https://doi.org/10.1002/hipo.22596

    • PubMed
    • Google Scholar
46. 1. Ng CW
    2. Elias GA
    3. Asem JSA
    4. Allen TA
    5. Fortin NJ

    (2018) Nonspatial sequence coding varies along the CA1 transverse axis

    Behavioural Brain Research 354:39–47.

    https://doi.org/10.1016/j.bbr.2017.10.015

    • PubMed
    • Google Scholar
47. Book

    1. O’Keefe J
    2. Nadel L

    (1978)

    The Hippocampus as a Cognitive Map

    Oxford: Clarendon Press.

    • Google Scholar
48. 1. Pesaran B
    2. Vinck M
    3. Einevoll GT
    4. Sirota A
    5. Fries P
    6. Siegel M
    7. Truccolo W
    8. Schroeder CE
    9. Srinivasan R

    (2018) Investigating large-scale brain dynamics using field potential recordings: analysis and interpretation

    Nature Neuroscience 21:903–919.

    https://doi.org/10.1038/s41593-018-0171-8

    • PubMed
    • Google Scholar
49. 1. Pogosyan A
    2. Gaynor LD
    3. Eusebio A
    4. Brown P

    (2009) Boosting cortical activity at Beta-band frequencies slows movement in humans

    Current Biology 19:1637–1641.

    https://doi.org/10.1016/j.cub.2009.07.074

    • PubMed
    • Google Scholar
50. 1. Rangel LM
    2. Chiba AA
    3. Quinn LK

    (2015) Theta and beta oscillatory dynamics in the dentate gyrus reveal a shift in network processing state during cue encounters

    Frontiers in Systems Neuroscience 9:96.

    https://doi.org/10.3389/fnsys.2015.00096

    • PubMed
    • Google Scholar
51. 1. Rangel LM
    2. Rueckemann JW
    3. Riviere PD
    4. Keefe KR
    5. Porter BS
    6. Heimbuch IS
    7. Budlong CH
    8. Eichenbaum H

    (2016) Rhythmic coordination of hippocampal neurons during associative memory processing

    eLife 5:e09849.

    https://doi.org/10.7554/eLife.09849

    • PubMed
    • Google Scholar
52. 1. Robinson NTM
    2. Priestley JB
    3. Rueckemann JW
    4. Garcia AD
    5. Smeglin VA
    6. Marino FA
    7. Eichenbaum H

    (2017) Medial Entorhinal Cortex Selectively Supports Temporal Coding by Hippocampal Neurons

    Neuron 94:677–688.

    https://doi.org/10.1016/j.neuron.2017.04.003

    • PubMed
    • Google Scholar
53. 1. Schmidt R
    2. Herrojo Ruiz M
    3. Kilavik BE
    4. Lundqvist M
    5. Starr PA
    6. Aron AR

    (2019) Beta Oscillations in Working Memory, Executive Control of Movement and Thought, and Sensorimotor Function

    The Journal of Neuroscience 39:8231–8238.

    https://doi.org/10.1523/JNEUROSCI.1163-19.2019

    • PubMed
    • Google Scholar
54. 1. Shahbaba B
    2. Li L
    3. Agostinelli F
    4. Saraf M
    5. Cooper KW
    6. Haghverdian D
    7. Elias GA
    8. Baldi P
    9. Fortin NJ

    (2022) Hippocampal ensembles represent sequential relationships among an extended sequence of nonspatial events

    Nature Communications 13:787.

    https://doi.org/10.1038/s41467-022-28057-6

    • PubMed
    • Google Scholar
55. 1. Stoffers D
    2. Berendse HW
    3. Deijen JB
    4. Wolters EC

    (2001) Motor perseveration is an early sign of Parkinson’s disease

    Neurology 57:2111–2113.

    https://doi.org/10.1212/wnl.57.11.2111

    • PubMed
    • Google Scholar
56. 1. Trimper JB
    2. Galloway CR
    3. Jones AC
    4. Mandi K
    5. Manns JR

    (2017) Gamma Oscillations in Rat Hippocampal Subregions Dentate Gyrus, CA3, CA1, and Subiculum Underlie Associative Memory Encoding

    Cell Reports 21:2419–2432.

    https://doi.org/10.1016/j.celrep.2017.10.123

    • PubMed
    • Google Scholar
57. 1. van Strien NM
    2. Cappaert NLM
    3. Witter MP

    (2009) The anatomy of memory: an interactive overview of the parahippocampal-hippocampal network

    Nature Reviews. Neuroscience 10:272–282.

    https://doi.org/10.1038/nrn2614

    • Google Scholar
58. 1. Weiss S
    2. Mueller HM

    (2012) “Too Many betas do not Spoil the Broth”: The Role of Beta Brain Oscillations in Language Processing

    Frontiers in Psychology 3:201.

    https://doi.org/10.3389/fpsyg.2012.00201

    • PubMed
    • Google Scholar
59. 1. Witter MP
    2. Doan TP
    3. Jacobsen B
    4. Nilssen ES
    5. Ohara S

    (2017) Architecture of the Entorhinal Cortex A Review of Entorhinal Anatomy in Rodents with Some Comparative Notes

    Frontiers in Systems Neuroscience 11:46.

    https://doi.org/10.3389/fnsys.2017.00046

    • PubMed
    • Google Scholar
60. 1. Zavala BA
    2. Jang AI
    3. Zaghloul KA

    (2017) Human subthalamic nucleus activity during non-motor decision making

    eLife 6:e31007.

    https://doi.org/10.7554/eLife.31007

    • PubMed
    • Google Scholar

Author details

1. Sandra Gattas

   1. Department of Electrical Engineering and Computer Science, University of California, Irvine, United States
   2. Center for the Neurobiology of Learning and Memory, University of California, Irvine, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Validation, Visualization, Writing - original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-1608-1469

2. Gabriel A Elias

   1. Center for the Neurobiology of Learning and Memory, University of California, Irvine, United States
   2. Department of Neurobiology and Behavior, University of California, Irvine, United States

   Contribution

   Data curation, Investigation, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-2339-7245

3. John Janecek

   1. Center for the Neurobiology of Learning and Memory, University of California, Irvine, United States
   2. Department of Neurobiology and Behavior, University of California, Irvine, United States

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

4. Michael A Yassa

   1. Center for the Neurobiology of Learning and Memory, University of California, Irvine, United States
   2. Department of Neurobiology and Behavior, University of California, Irvine, United States

   Contribution

   Funding acquisition, Investigation, Supervision, Writing – review and editing

   For correspondence

   michael.yassa@uci.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8635-1498

5. Norbert J Fortin

   1. Center for the Neurobiology of Learning and Memory, University of California, Irvine, United States
   2. Department of Neurobiology and Behavior, University of California, Irvine, United States

   Contribution

   Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – review and editing

   For correspondence

   norbert.fortin@uci.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-6793-6984

• 800

  views

• 142

  downloads

• 4

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.