Beta bursting in the retrosplenial cortex is a neurophysiological correlate of environmental novelty which is disrupted in a mouse model of Alzheimer’s disease

The retrosplenial cortex (RSC) plays a significant role in spatial learning and memory, and is functionally disrupted in the early stages of Alzheimer’s disease. In order to investigate neurophysiological correlates of spatial learning and memory in this region we employed in vivo electrophysiology in awake, behaving mice, comparing neural activity between wild-type and J20 mice, a mouse model of Alzheimer’s disease-associated amyloidopathy. To determine the response of the RSC to environmental novelty local field potentials were recorded while mice explored novel and familiar recording arenas. In familiar environments we detected short, phasic bursts of beta (20-30 Hz) oscillations (beta bursts) which arose at a low but steady rate. Exposure to a novel environment rapidly initiated a dramatic increase in the rate, size and duration of beta bursts. Additionally, theta-beta cross-frequency coupling was significantly higher during novelty, and spiking of neurons in the RSC was significantly enhanced during beta bursts. Finally, aberrant beta bursting was seen in J20 mice, including increased beta bursting during novelty and familiarity, yet a loss of coupling between beta bursts and spiking activity. These findings, support the concept that beta bursting may be responsible for the activation and reactivation of neuronal ensembles underpinning the formation and maintenance of cortical representations, and that disruptions to this activity in J20 mice may underlie cognitive impairments seen in these animals.


18
The retrosplenial cortex (RSC) is considered to play a critical role in spatial learning and memory. 19 Damage to this region results in severe impairments in navigation and landmark processing (see 20 Mitchell et al., 2018 for review). There is a large body of experimental evidence suggesting the 21 retrosplenial cortex is involved in the encoding, retrieval and consolidation of spatial and contextual 22 memory (see Todd  score exceeded 2 standard deviations from the mean amplitude were detected, as were the 137 corresponding "edges" of these epochs, where the signal magnitude surpassed 1 standard deviation 138 either side of the 2 standard deviation threshold. This was done in order to capture the time-course 139 of these high beta amplitude epochs. Events that did not persist longer than a minimum duration of 140 150 ms (i.e. fewer than 3 oscillation cycles) were discarded. Furthermore, due to the sensitivity of this 141 method to large, amplitude noise artefacts, any event whose peak amplitude exceeded three scaled 142 median absolute deviations from the median of the events detected in that session were discarded as 143 well. These remaining events were then considered beta-bursts. The duration and peak magnitude of 144 each burst was calculated, as well as the distribution and total number of bursts in the session. 145

Phase-Amplitude Coupling
individually for each pair of phase and amplitude frequencies. Modulation index was calculated as 148 described by Canolty et al. (2006), with modification and vectorisation of some of the MATLAB code, 149 for phase frequencies in bins of 0.25 Hz from 2 to 12 Hz, and for amplitude frequencies in bins of 2 Hz 150 from 10 to 100 Hz. For each pair, local field potentials were filtered in the phase frequency band and 151 the amplitude frequency band, after which the instantaneous phase and amplitude of each filtered 152 signal was calculated, respectively, using the Hilbert transform. Subsequently, modulation index (MI) 153 was calculated, but in order to attempt to reduce the possibility of spurious coupling, this was 154 normalised through the use of 10 surrogates, created by time shifting the data by a random amount 155 (between 1 and 59 seconds). In order to smooth the resulting comodulograms, the data matrix was 156 linearly interpolated in both dimensions by a factor of 2. 157

158
Due to the distance between adjacent channels on the recording probe (100 µm) it is highly unlikely 159 that activity of a single neuron would appear on multiple channels. Consequently, each channel was 160 treated as an individual multi-unit. Raw local field potentials were first common average-referenced, 161 using a mean of the signals from all other 15 channels, then filtered in the range of 500-14250 Hz, in 162 order to isolate the spiking frequency band. Spikes were detected as peaks that crossed a threshold 163 given by the median of the absolute voltage values of the signal, multiplied by 0.6745, as suggested 164 by Quiroga, Nadasdy and Ben-Shaul (2004), and had a minimum separation of 0.5 ms. In order to 165 investigate multi-unit activity during beta bursts, bursts were detected as previously mentioned, and 166 bursts that occurred within a second of each other were discarded, to remove overlapping segments. 167 A single peri-burst histogram was created for each channel by taking the total number of spikes in 20 168 ms time bins from 1 second before burst onset, to 1 second after, for all beta bursts. Each histogram 169 was then normalised by dividing the count in each bin by the total number of spikes in all bins,

177
All statistical analysis was performed in MATLAB. Fourteen mice in total were used in this study, 6 178 wild-type and 8 J20, with each mouse undergoing a total of ten recording sessions (5 days, 2 sessions 179 per day). Unfortunately, the local field potential data from Day 3 session 1 (i.e. session 3a) was 180 corrupted for a single wild-type mouse, and therefore data for this mouse from this session was 181 omitted from all figure making and statistics. Therefore the n numbers for all statistics are (wild-type: 182 n = 6 (except from Day3a where n = 5), J20: n = 8). All statistics, unless stated otherwise, were 183 performed using a two-way ANOVA, with genotype (wild-type/J20) and novelty (novel/familiar) as 184 factors. It is important to note that due to the experimental design of our Novel/Familiar environment 185 task, there were multiple novel and familiar sessions (2 novel, and 8 familiar). All sessions were either 186 classified as novel or familiar and analysed accordingly. Following a significant main effect or 187 interaction, Bonferroni-corrected multiple comparisons was performed, to investigate pairwise 188 differences between different levels of either factor. 189

190
Upon completion of the experiments, mice were killed using an overdose of sodium pentobarbital 191 (Euthetal), and an isolated stimulator was used to produce electrolytic lesions at the recording sites. 192 Mice were then transcardially perfused with 40% paraformaldehyde (PFA), and their brains were 193 extracted and stored in PFA for 24 hours, after which they were transferred to phosphate-buffered count of these probes, as well as their linear geometry, it was possible to account for small differences 199 in the depth of each probe by selecting channels of similar depths across different probes. This 200 resulted in reduced variability between animals in a range of neurophysiological measures. 201

202
To investigate neurophysiological correlates of spatial learning and memory in the retrosplenial cortex 203 (RSC), local field potentials were recorded from across the entire dorsoventral axis of the RSC, while 204 animals underwent a novel/familiar environment task. The RSC is made up of two distinct subdivisions: 205 dysgranular (RSCdg), and granular (RSCg). While these regions are strongly interconnected, the 206 neuroanatomical connectivity of these regions has been shown to differ (van Groen and Wyss, 1992; 207 Van Groen and Wyss, 2003a, 2003b), therefore it is possible that the functional neurophysiology may 208 vary as well, especially during a behavioural paradigm such as this, where spatial learning and memory 209 processes may be stimulated. Due to the anatomical positioning of these subdivisions in rodents, it is 210 possible to record from both RSCdg and RSCg at once, using a single, vertical silicon probe ( Figure 1C). 211 Therefore for this study, our analyses were performed for both subdivisions. We found very little 212 difference between the electrophysiological activities seen in the two subregions. Furthermore, any 213 changes seen in J20 mice were generally common to both subregions, with marginally greater effects 214 in RSCg. For the sake of conciseness, we have decided to only show the data from RSCg in this paper. 215

216
Local field potentials from RSCg show a clear peak in theta frequency band (5-12 Hz) throughout 217 recording sessions (Fig. 2a). In order to investigate any changes in oscillatory activity in RSCg during 218 expanded power spectrogram in (Fig. 3a). This can also be seen clearly in beta-filtered local field 245 potentials, where these periods of high beta amplitude intersperse an otherwise very low amplitude 246 oscillation. In order to understand the timescale and frequency domains of these events, wavelet 247 analysis was used to investigate them further. As exemplified in (Fig. 3C), these individual events were 248 short in duration, and peaked in the 20-30 Hz, beta band. 249

250
In order to investigate this phasic beta activity in more depth, an algorithm was written to detect and 251 analyse these "beta bursts"; the basis of this algorithm is illustrated in (Fig. 4a). Once all putative bursts 252 have been detected, the duration and magnitude of these beta bursts was calculated (Fig. 4a). With 253 these transient epochs of high beta power now classified as discrete beta bursts, it is possible to 254 compare this beta activity between sessions. Overall, there were significantly more beta bursts 255 detected during novel sessions compared to familiar sessions (Main Effect Novelty -F(1,135) = 74, p = 256 1.73e-14, 2-way ANOVA). As shown in (Fig. 4b), there were significantly more beta bursts detected 257 during novelty, for wild-type (Nov: 33.7 ± 2.42; Fam: 21.4 ± 1.22; p = 7.59e-5) and J20 mice (Nov: 56.3 258 ± 2.1; Fam: 37.8 ± 1.05; p = 4.83e-12). Furthermore, on average the number of beta bursts detected 259 was significantly higher in J20 mice (Main Effect Genotype -F(1,135) = 118, p = 3.45e-20, 2-way 260 ANOVA). Furthermore, it is possible to investigate the distribution of beta bursts within sessions. As 261 shown in (Fig. 4c), during familiar sessions the rate of beta busting was reasonably steady, as indicated 262 by the linear relationship between time and burst number shown in the cumulative frequency plot, 263 for both wild-type and J20 mice. During novel sessions, however, there was a high rate of beta bursting 264 during the first 1-3 minutes of the session, which gradually decreased over time to a steady rate. The 265 rate of beta bursting was significantly higher in J20 mice during familiar sessions, and during the initial 266 and final part of novel sessions. 267 interaction between the effects of genotype and novelty on beta burst duration (F(1,135) = 8.04, p = 274 0.005, 2-way ANOVA). As shown in (Fig.4e), beta bursts were significantly longer in duration during 275 novel sessions for both wild-type (Nov: 192 ± 2.1; Fam: 176 ± 1.1; p =3.32e-9) and J20 mice (Nov: 189 276 ± 1.8; Fam: 182 ± 0.9; p = 0.005). 277 if their amplitude varies greatly over time. The amplitude of high frequency oscillations such as gamma 280 may be modulated by the phase of low frequency oscillations such as theta (Canolty et al., 2006). This 281 interaction is generally thought to allow slow, large amplitude oscillations to coordinate faster, small 282 amplitude local oscillations. For this reason, it was of interest for us to investigate whether the 283 amplitude of beta oscillations was coupled to the phase of theta oscillations, an increase in which may 284 underlie the increased beta bursting activity seen during novelty. As shown in (Fig. 5a), phase-285 amplitude coupling efficacy was calculated for a range of phase and amplitude frequencies, and the 286 effect of novelty and genotype determined. The strength of phase-amplitude coupling was quantified 287 for theta-alpha, theta-beta and theta-gamma coupling for each session (Fig. 5b). There were 288 significant interactions between the effects of genotype and novelty for theta-alpha coupling (F(1,135) 289 = 12.8, p = 4.72e-4) and theta-beta coupling (F(1,135) = 17.7, p = 4.73e-5, 2-way ANOVA). Theta-alpha 290 coupling was significantly higher in novel sessions for wild-type (Nov: 2.59 ± 0.15; Fam: 1.6 ± 0.07; p = 291 2.4e-7) but not J20 mice (Nov: 2.2 ± 0.13; Fam: 1.98 ± 0.06; p = 1). Theta-beta coupling was also physiologically and behaviourally relevant part of the session, this analysis was performed for the first 297 minute of each session. When the same analysis was performed on the last minute of each session, 298

Phase-amplitude Coupling
there was no effect of novelty on coupling in any band for either genotype (data not shown). 299

300
In order to determine whether beta bursting was associated with a change in neuronal firing, multi-301 unit activity was investigated. Due to the linear geometry of the silicon probes, and the 100 µm 302 distance between channels, it was not possible to reliably identify single unit activity, as activity from 303 a single neuron was unlikely to appear on multiple channels, limiting spatiotemporal clustering 304 methods such as those enabled by tetrodes or higher density silicon probes. Therefore, spikes 305 appearing on a single channel could be from one or more nearby neurons. This, however, does mean 306 that it is possible to treat each individual probe channel as a single multi-unit, to facilitate investigation 307 of the relationship between neuronal spiking activity and beta bursting. As shown in the left panel of 308 ( Fig. 6a), individual spike waveforms can be readily discerned, and these spike waveforms are similar 309 in wild-type (black) and J20 (green) mice. Furthermore, there was a trend towards higher multi-unit 310 firing rate in J20 mice compared to wild-type mice (WT: 12.9 Hz ± 4.9; J20: 33.5 Hz ± 7.3; t(12) = -2.18, 311 p = 0.05; unpaired t-test, Fig. 6a, right). The average beta amplitude during beta bursts is shown in 312 (Fig. 6b), averaged across all bursts with non-overlapping time segments. Beta bursts in both 313 genotypes are associated with a brief, monophasic increase in beta amplitude that lasts no more than 314 200 ms on average. Finally, (Fig. 6c) shows peri-event time histograms for spike rate during beta 315 bursts, as a Z score from the pre-burst baseline (left of the dotted line). In order to investigate 316 statistically significant changes in spike rate during bursts, the maximum z scored spike rate was and compared to the mean pre-burst spike rate (0 due to z scoring of spike rate to baseline) using a 319 one-sample t-test. Beta bursting in the RSCg of wild-type mice was associated with a significant 320 increase in spike rate during beta bursts (Z-scored spike rate from baseline: 2.24 ± 0.46, t(5) = 4.86, p 321 = 0.005, one-sample t-test; Figure 6c, left). Conversely there was no significant increase in spike rate 322 during beta bursts in J20 mice (Z-scored spike rate from baseline: 0.78 ± 0.39, t(7) = 1.98, p = 0.09, 323 one-sample t-test; Figure 6c, right). The difference between spike rate during beta bursts in wild-type 324 and J20 mice, as determined by a two-sample t-test, was significant (t(12) = 2.4, p = 0.03, two-sample 325 t-test). These data indicate that beta bursts are closely coupled to neuronal spiking in RSCg in wild-326 type mice, and that this relationship is effectively uncoupled in J20 mice. 327

328
In this study we attempted to identify neurophysiological correlates of environmental novelty in the 329 mouse retrosplenial cortex (RSC), and investigate how these may be affected by amyloid pathology. 330 We observed phasic increases in the amplitude of beta frequency neuronal oscillations, termed beta 331 bursts, which occurred more frequently and with larger amplitude during novelty, and were positively 332 correlated with neuronal spiking. A number of aberrant neurophysiological changes were seen in the 333 RSC in J20 mice. Alpha, beta and low gamma power were significantly increased, and increases in beta 334 bursting activity were seen during both novelty and familiarity. Beta bursts were more frequent, and 335 larger in magnitude, yet the coupling of beta bursts to spiking activity was lost, suggesting a functional 336 uncoupling of beta bursting with local neuronal activity. Finally, theta-beta phase-amplitude coupling 337 was also disrupted, resulting in a loss of an effect of novelty on this activity. These results together 338 indicate that beta bursting activity is a neurophysiological correlate of environmental novelty in the 339 RSC, which is disrupted in J20 mice. 340 Numerous studies have noted changes in beta activity in a range of brain regions, during a variety of 341 behaviours (see Spitzer and Haegens, 2017 for review). It is important to note that due to variability explored a novel environment, which persisted for around a minute, before returning to a lower level. 349 The authors concluded that these oscillations may be a "dynamic state that facilitates the formation Furthermore, they found that administration of an amnestic agent, namely haloperidol, resulted in a 356 similar increased beta activity upon re-exposure to previously encountered objects, suggesting they 357 had been "forgotten" and were therefore novel again. This further reinforces the idea that 358 hippocampal beta activity is related to novelty, and extends the previous work by demonstrating that 359 hippocampal-dependent novel object recognition can also elicit beta oscillations. Subsequently, 360 França, Borgegius and Cohen (2020) investigated novelty-associated beta bursting in a larger 361 hippocampal novelty circuit, by simultaneously recording from hippocampus, prefrontal cortex and 362 parietal cortex during environmental and object novelty. Novelty-associated increases in beta power 363 were seen in the prefrontal cortex during environmental novelty, and authors demonstrated 364 significant phase-amplitude coupling of delta and theta to beta oscillations, which were increased in 365 novelty. Similarly, in the RSC we see strong coupling between theta phase and beta amplitude, which 366 is significantly higher during novelty, but only in wild-type mice. Others have noted theta-beta PAC in 367 humans as well, both in the hippocampus during a working memory task (Axmacher et al., 2010), and This is despite Berke et al. (2008) noting that beta appears as pulses, and a brief mention of burst 371 detection and characterisation by França et al. (2014). As demonstrated in this study, novelty-372 associated beta oscillations in the RSC conform well to a model of discrete bursts, where their rate, 373 magnitude and duration can vary depending on environmental novelty. Due to the use of averaging 374 across trials or analysis spanning long temporal segments, the phasic nature of transient oscillatory 375 events can be easily lost. Furthermore, in the somatosensory cortex, beta synchronicity appears in 376 short events in both mice and humans; the features of which, such as duration and frequency range, 377 are highly conserved across tasks and species (Shin et al., 2017). 378 Beta oscillations have long been associated with motor activity and sensory processing, and a large 379 body of work has also noted changes in beta activity in a range of brain regions during other cognitive 380 tasks (see Engel and Fries, 2010 for review). This gave rise to the hypothesis that the unifying function 381 of beta oscillatory activity in these different regions was the maintenance of the "status-quo", be it 382 the current motor state, sensory stimulus or cognitive set (Engel and Fries, 2010). This theory would 383 suggest that, beta activity would be decreased during novelty, and increased during familiarity. As we 384 have shown, this is not the case. While steady and persistent beta bursting during familiarity may 385 support the maintenance of the contextual "status-quo", in this case the environment, this theory 386 does not reconcile the significant increases in beta activity that occur during novelty. range of brain regions, and phasic increases in beta power during working memory maintenance may 391 represent reactivation of encoded information (Spitzer and Haegens, 2017). Supporting this is a study 392 in which the authors employed transcranial magnetic stimulation to activate a currently unattended 393 Figure 1. Experimental Design A. Diagrams of the recording arenas used for this study. Both are roughly equal sized, one is square, with black and white stripes along the walls and floor (left) and the other is cylindrical with plain brown floor and walls. B. Experimental procedure for the novel/familiar environment task. A mouse is placed in one of the recording arenas for two 15 minute sessions, referred to as sessions A and B, with a 15 minute break in their home cage between the two sessions. This is repeated in the same arena for 4 consecutive days, after which the arena is switched for the 5 th and final day. C. Single shank, 16 channel silicon probe electrodes were implanted in the retrosplenial cortex (green), so that they spanned the dysgranular (upper green section) and granular (lower green section) subregions. In order to verify the location of the electrodes, electrolytic lesions were made prior to perfusion, and slices were histologically prepared using Cresyl Violet stain. An example is shown (right). Figure 2. Beta (20-30 Hz) power is significantly higher during novelty in the granular retrosplenial cortex in wild-type and J20 mice. A. Example power spectrogram for an entire novel session in a wild-type mouse. B. Average power spectra for the entire 15 minutes of all novel and familiar sessions, for wild-type and J20 mice. Beta power was significantly higher during novelty in J20 (p = 4.7e-4) but not wild-type mice. When compared to WT power in the alpha, beta, low gamma and high gamma bands were significantly higher overall in J20 mice (p = 8.46e-6, p = 1.01e-32, p = 9.56e-10, 2.3e-4 respectively), whereas power in the delta and theta band were significantly lower (p = 0.03, p = 0.006 respectively). C. Example power spectrogram shown in A, expanded to show the first 60 seconds of the session. Short epochs of increased power in the 20-40 Hz range can be seen. D. Average power spectra for the first minute of all novel and familiar sessions, for wild-type and J20 mice. Beta and low-gamma power were significantly higher during novelty, for both wildtype (p = 5.47e-8, p = 3.62e-5 respectively) and J20 mice (p = 3.59e-8, p = 0.002 respectively). Alpha, beta, low gamma and high gamma power were significantly higher overall in J20 mice (p = 2.47e-9, p = 1.1e-21, p = 2.52e-4, p = 4.65e-4 respectively). (Data shown as mean ± SEM, WT: n = 6, J20: n = 8).  . Beta bursting activity in the granular retrosplenial cortex (RSCg) is highly associated with novelty, and dysregulated in J20 mice. A. Diagram illustrating how beta bursts were detected, as well as how the magnitude and duration of these events were calculated. B. Graph showing the average number of beta bursts detected in RSCg in each session, for wild-type (black) and J20 mice (green). Novel sessions Day1a and Day5a are highlighted in blue for clarity. There were significantly more beta bursts in novel sessions as compared to familiar sessions, for both wild-type (p = 7.59e-5) and J20 mice (p = 4.83e-12). C. Cumulative frequency graphs of number of bursts detected in novel and familiar sessions, for wild-type and J20 mice, showing the time course of bursting activity within sessions. While beta bursting occurred monotonically during familiar sessions, during the first 2-3 minutes of a novel session, beta bursting was substantially increased. D. Graph showing the average magnitude of beta bursts in RSCg in each session, for wild-type and J20 mice. Beta bursts were significantly larger in magnitude in novel sessions, for wild-type (p = 2.88e-5) and J20 mice (p = 5.16e-6). Beta bursts were also, on average, significantly larger in magnitude in J20 mice (p = 2.97e-22). E. Graph showing the average duration of beta bursts in RSCg in each session, for wild-type and J20 mice. Beta bursts were significantly longer in duration in novel sessions, for wildtype (p = 3.32e-9) and J20 mice (p = 0.005). (Data shown as mean ± SEM, WT: n = 6, J20: n = 8). Figure 5. Theta-alpha and theta-beta phase-amplitude coupling are increased during novelty in the granular retrosplenial cortex (RSCg). A. Average comodulograms showing the strength of crossfrequency phase-amplitude coupling in RSCg during the first minute of novel and familiar sessions, for wild-type and J20 mice. Note the presence of three peaks in the first comodulogram, in the theta-alpha, theta-beta and theta-gamma ranges (the boundaries of which are denoted by the dotted lines). B. Average MI in the theta-alpha (left), theta-beta (center) and theta-gamma ranges (right), for each session, for wild-type (black) and J20 mice (green). Novel sessions Day1a and Day5a are highlighted in blue for clarity. Theta-alpha and theta-beta coupling were significantly higher in novel sessions for wild-type mice (p = 2.4e-7, p = 1.04e-6 respectively), but not J20 mice. (Data shown as mean ± SEM, WT: n = 6, J20: n = 8). Figure 6. Spiking activity in RSCg is coupled to beta bursting in wild-type mice, but disrupted in J20 mice. A. Average spike waveforms for multi-unit activity in wild-type (black) and J20 (green) mice (left) and graph of average firing rate for detected multi-units across all sessions (right). There was a trend towards increased multi-unit firing rate in J20 mice compared to wild-type mice (p = 0.052, unpaired t-test). B. Graphs showing beta amplitude over time for beta bursts, time locked to the onset of the burst (dotted line), and averaged across all detected bursts, for wild-type mice (left) and J20 mice (right). Beta bursting was associated with a monophasic increase in beta amplitude that returns to baseline after around 250 ms. C. Peri-event histograms showing multi-unit activity spike rate during beta bursts, for wild-type (left) and J20 mice (right). Data is shown as Z score from baseline (pre-burst epoch), and averaged across all beta bursts with non-overlapping time segments. Dotted vertical line denotes the burst onset, while the solid horizontal line is shown to indicate the baseline of zero. Spike rate significantly increased during bursts in wild-type mice (p = 0.005), but not in J20 mice (p = 0.09).