Neurobiology of Aging 79 (2019) 142e151 Contents lists available at ScienceDirect Neurobiology of Aging journal homepage: www.elsevier.com/locate/neuaging Altered slow (<1 Hz) and fast (beta and gamma) neocortical oscillations in the 3xTg-AD mouse model of Alzheimer’s disease under anesthesia Patricia Castano-Prat a, Lorena Perez-Mendez a, Maria Perez-Zabalza a, Coral Sanfeliu b, Lydia Giménez-Llort c, Maria V. Sanchez-Vives a, d, * a Systems Neuroscience, IDIBAPS (Institut d’Investigacions Biomèdiques August Pi i Sunyer), Barcelona, Spain b Biomedical Research Institute of Barcelona (IIBB), Consejo Superior de Investigaciones Científicas (CSIC), IDIBAPS and CIBERESP, Barcelona, Spain c Institut de Neurociènces and Department of Psychiatry and Forensic Medicine, School of Medicine, Universitat Autònoma de Barcelona (UAB), Barcelona, Spain d ICREA, Barcelona, Spain a r t i c l e i n f o a b s t r a c t Article history: The 3xTg-AD mouse model reproduces the main features associated with the etiology of familial Alz- Received 11 April 2018 heimer’s disease (AD). To investigate whether these features imply functional cortical network alter- Received in revised form 24 January 2019 ations and their evolution with age, we studied spontaneous slow oscillations, activity that integrates Accepted 9 February 2019 cellular and network properties. We quantified different parameters of the emergent slow oscil- Available online 13 March 2019 lationsdalternating Up and Down statesdand of the embedded beta-gamma rhythms of 3xTg-AD and wild-type mice at 7 and 20 months of age. Most group differences occurred at 20 months of age: 3xTg-AD Keywords: mice presented lower oscillatory frequency, higher cycle variability, and reduced relative (Up/Down) Aging Alzheimer disease firing rate with respect to controls. The high-frequency analysis revealed a shift toward lower frequencies Gamma oscillations in older 3xTg-AD animals, reminiscent of one of the electroencephalography hallmarks of patients with Gamma rhythm AD. This first systematic characterization of the cortical emergent rhythms in 3xTg-AD strain provides Slow oscillations insights into the network mechanisms underlying associated network activity alterations. Up states Ó 2019 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND Cerebral cortex license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Transgenic model Excitability 1. Introduction manner that resembles that in human AD brain (LaFerla and Oddo, 2005; Oddo et al., 2003a,b). At 7 months of age, 3xTg-AD Alzheimer’s disease (AD) is a primary neurodegenerative dis- mice present intracellular Ab immunoreactivity in the neocortex order that causes a progressive impairment of memory and other (especially in the prefrontal cortex), the CA1 subfield of the hip- cognitive functions, and represents the leading cause of dementia pocampus, and the amygdala, as well as long-term potentiation among the elderly. The principal neuropathological hallmarks of AD deficits and altered basal synaptic transmission (Oddo et al., are the abnormal accumulation of the amyloid-b (Ab) peptide 2003b). At the behavioral level, these changes result in altered forming Ab plaques, the hyperphosphorylation of the microtubule- fear conditioning, impaired learning, and memory retention in the associated protein taudknown as neurofibrillary tanglesdthe loss Morris water maze task (Billings et al., 2005) and poor performance of neurites and synapses, and cellular death at advanced stages of in object recognition (Filali et al., 2012). By 20 months of age, the disease (for a review, see Querfurth and LaFerla, 2010). Most of extracellular Ab plaques extend throughout the neocortex (espe- the mouse models of AD mimic only some of the aforementioned cially in layers IV and V of the prefrontal cortex) and the CA1 area. features of the disease (Wong et al., 2002). The triple-transgenic At this advanced stage of the disease, neurofibrillary tangles are mouse model of AD (3xTg-AD), on the other hand, develops Ab also evident in the hippocampus and the neocortex (Oddo et al., and tau pathology in a progressive, temporal- and region-specific 2003b; Sy et al., 2011), alongside a considerable decrease in den- dritic spine density (Bittner et al., 2010) and a significant reduction * Corresponding author at: IDIBAPS, C/ Rosselló 149-153, 08036 Barcelona, Spain. of the cerebral metabolism in cortical, subcortical, cerebellar, and Tel.: 34 932 275 400; fax: þ34 932 271 890. brainstem regions (Nicholson et al., 2010). E-mail address: msanche3@clinic.ub.es (M.V. Sanchez-Vives). 0197-4580/Ó 2019 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). https://doi.org/10.1016/j.neurobiolaging.2019.02.009 P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 143 These neuropathological and behavioral AD-like hallmarks in the (40 mg$kg1$hour1) was delivered subcutaneously with a pump to 3xTg-AD mouse model are likely to be accompanied by disturbances maintain a constant level of anesthesia (Castano-Prat et al., 2017; in cortical network activity. One way to assess these disturbances is Ruiz-Mejias et al., 2011). Body temperature was maintained at 37 by studying the spontaneously emergent activity of the cortical  Ce38  C throughout the experiment. Following Paxinos and network, which integrates the neuronal intrinsic and synaptic Franklin (2008), bilateral craniotomies were made at 4 sites in each properties and those of connectivity. Cortical slow oscillations (SO) mouse: AP 2.3 mm from bregma, L 0.4 mm (prelimbic cortex [PrL] or occur during natural sleep and under anesthesia. They consist in a medial prefrontal cortex); AP 0.5 mm, L 1.5 mm (primary motor low-frequency (1 Hz) rhythmic pattern in which cortical neurons cortex, M1); AP 1.5 mm, L 2.5 mm (primary somatosensory cortex, fluctuate between active periods of neuronal firing and synaptic S1); and AP 2.5 mm, L 1.5 mm (primary visual cortex, V1). Re- activity (Up states) and relatively inactive periods (Down states) cordings to survey local neuronal activity were obtained from deep (Steriade et al.,1993). Up states are generated by the cortical network layers (0.4 mm apart from the midline and 1.1e1.2 mm deep in PrL, itself, and are maintained by the network’s recurrent excitatory and 0.9e1.0 mm deep in M1, 0.8e0.9 mm deep in S1, and 0.8e0.9 mm inhibitory synaptic connections. Furthermore, the activity pattern deep in V1) by means of 1e2 MU tungsten microelectrodes (tip size emerging during the Up states replicates and predicts some of the less than 10 mm) insulated with a plastic coating except for the tip features during periods of cortical activation such as during rapid (FHC, Bowdoin, ME, USA). Spontaneous local field potential (LFP) eye movement sleep and wakefulness (Destexhe et al., 2007; recordings were obtained sequentially from each cortical area and Steriade et al., 1996; Steriade and Timofeev, 2003). simultaneously from left and right hemispheres. LFP recordings SO rhythmic activity can provide relevant information about the provide information about the local neuronal populationdwithin network parameters that are compromised in the cerebral cortex of 250 mm according to Katzner et al. (2009)dincluding synaptic po- 3xTg-AD mice and also bring useful insights into the underlying tentials and, when high-pass-filtered, the multiunit activity (MUA). alterations at the cellular and synaptic levels (Neske, 2016; Multichannel LFP recordings of spontaneous activity from different Sanchez-Vives and Mattia, 2014). Impaired network synchroniza- areas therefore inform about the state of different neuronal ensem- tion has been reported in the hippocampus of 3xTg-AD mice (Akay bles, their firing rates, rhythmic activity, and synchronization at et al., 2009; Squirrell, 2015), but cortical network disturbances have different frequency bands. All these parameters allow a detailed not been assessed in this strain. Thus, the aim of this study was to comparison between control and 3xTg-AD mice and an estimation of investigate whether 3xTg-AD mice presented any disturbances in properties such as excitability and excitatory/inhibitory balance. In cortical network activity at 2 different ages, to evaluate also changes each animal, the 4 cortical areas were recorded in a counterbalanced associated with aging. order to avoid any influence of anesthesia levels, even though a constant level of anesthesia was maintained throughout the exper- 2. Methods iment. The signal was amplified with a multichannel system (Multi Channel Systems), digitized at 20 kHz with a CED acquisition board 2.1. Animals and acquired with Spike2 software (Cambridge Electronic Design) unfiltered. The 3xTg-AD mouse strain was genetically engineered at the University of California Irvine to express familial AD mutations of 2.3. Data analysis amyloid precursor protein (APPSwe), presenilin-1 (PS1M146V), and tau (tauP301L) (Oddo et al., 2003a,b). The mice included in this study To assess cortical network activity in 3xTg-AD and control mice, were bred at the University of Barcelona and the Autonomous Uni- we analyzed and compared several parameters of the SO: its mean versity of Barcelona and were provided by Drs Lydia Giménez-Llort frequency, Up and Down state duration, the coefficient of variation and Coral Sanfeliu (Gimenez-Llort et al., 2007). To determine whether (CV) of the Up-Down state cycle, and the population firing rate. 3xTg-AD mice present disturbances in cortical network activity and to Detection of Up and Down states from the recorded signals was assess its evolution with aging, extracellular recordings were obtained based on 3 main fingerprints of the Up states: the slow LFP from 7-month-old (3xTg n ¼ 18; wild-type [Wt] n ¼ 14) and 20- deflection, the gamma rhythm and the neuronal firing. These 3 month-old (3xTg n ¼ 5; Wt n ¼ 8) adult mice of either sex. Mice features are reflected in 3 different time series: (1) the SO envelope; were cared for and treated in accordance with Spanish regulatory laws (2) the envelope of the variance of the gamma-filtered LFP (BOE-A-2013-6271), which comply with the European Union guide- (Mukovski et al., 2007); and (3) the estimation of MUA (Reig et al., lines on protection of vertebrates used for experimentation (Directive 2010; Sanchez-Vives et al., 2010). From each LFP, we obtained a 2010/63/EU of the European Parliament and the Council of 22 highly processed time series as a linear combination of these 3 September 2010). All experiments were approved by the Ethics features. The contribution of each one was weighted by principal Committee at the Hospital Clinic de Barcelona (Spain). Mice were kept component analysis by their projection into the first principal axis under standard conditions (room temperature at 23  1  C, 12:12-h obtained in a principal component analysis. As the 3 signals light-dark cycle, lights on at 08:00 a.m.), with food (A04, Harlan, correspond to 3 different frequency bands, this method is very Spain) and water available ad libitum throughout the study. robust against colored noise or band-limited electrode malfunction. Up and Down states were singled out by setting a threshold in this 2.2. In vivo extracellular recordings highly processed time series and the detection was done by an al- gorithm. This method has been widely used previously (e.g., Ruiz- Anesthesia was induced by means of intraperitoneal injection of Mejias et al., 2011, 2016; D’Andola et al., 2018). Parameters of the ketamine (100 mg/kg) and medetomidine (1.3 mg/kg). Atropine SO were computed from bilateral recordings in the 4 recorded (0.3 mg/kg) and methylprednisolone (30 mg/kg) were administered areas. The frequency of the SO was the inverse of the duration of the subcutaneously to prevent respiratory secretions and inflammation, whole Up-Down cycle. The CV of the SO frequency was the ratio respectively. A tracheotomy was performed to increase stability between the standard deviation and the mean of the SO frequency. during the recordings. After this procedure, the mouse was placed in The firing rate was based on the MUA signal. The MUA was a stereotaxic frame and air enriched with oxygen was delivered to estimated from the extracellular recordings as the power change in the mouse through a thin silicon tube placed at 0.5e1 cm from the the Fourier components at frequencies between 200 and 1500 Hz in tracheal cannula. Continuous infusion of ketamine 5-ms windows (Reig et al., 2010; Sanchez-Vives et al., 2010). We 144 P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 assumed that the spectrum within this frequency band provides a differences between right and left hemispheres, data from both good estimate of the population firing rate because Fourier com- hemispheres were averaged for each area and subject. Comparisons ponents at high frequencies have densities proportional to the between groups and ages were performed by means of repeated spiking activity of the involved neurons (Mattia and Del Giudice, measures two-way ANOVA for each parameter, setting the cortical 2002). To obtain MUA time series (logMUA), MUA values were area (PrL, M1, S1, and V1) as the within-subjects variable, and age (7 logarithmically scaled to balance the large fluctuations of the and 20 months of age) and group (Wt and 3xTg-AD) as between- nearby spikes. The absolute values of firing rates during Up and subjects variables. Post hoc tests were performed with Student’s t- Down states were computed as the mean of logMUA values during tests for independent samples, corrected for multiple comparisons homologous periods. The relative firing rate was the maximum using the Bonferroni method (corrected a ¼ 0.0125). All the analyses value of the logMUA averaged waveform after the Down-to-Up were implemented in MATLAB (the MathWorks, Natick, MA, USA). transition, after being normalized to zero during the Down states. Finally, to analyze the fast components of the SO, we carried out Welch’s power spectrum density analysis during Up and Down 3. Results states, with 50% overlapped windows of 5000 samples. To highlight and compare peaks of neuronal synchronization over the intrinsic Spontaneous rhythmic activity generated in the deep layers of exponential decay of the signal, we computed the excess power, different cortical regions (PrL, M1, S1, and V1; Fig. 1A) of 3xTg-AD defined as the power ratio between power and the “1/f” decay. and control mice was recorded under deep anesthesia at 7 and 20 months of age. Under these conditions, SO activity (1 Hz) with alternating periods of intense synaptic activity and neuronal firing 2.4. Statistical analysis (Up states) and rather quiescent periods (Down states) was generated in the thalamocortical network (Fig. 1B, top) (Steriade Comparisons between right and left hemispheres were performed et al., 1993). Fast rhythms (beta-gamma, 15e100 Hz) were also separately for each cortical area and parameter with the Student’s t- detected (Fig. 1B, bottom), mainly confined to the Up states test for dependent samples, corrected for multiple comparisons with (Compte et al., 2008; Steriade et al., 1996). Following previous the Bonferroni method (corrected a ¼ 0.0125). As we did not find any studies (Castano-Prat et al., 2017; Ruiz-Mejias et al., 2011; Fig. 1. Spontaneous oscillatory activity in the 3xTg-AD mouse. (A): Extracellular activity in deep layers was recorded simultaneously in both hemispheres in each cortical area. (B): Top: Raw local field potential (LFP) from the right PrL cortex of a wild-type (Wt) mouse showing slow oscillations (SO) with alternating Up (active) and Down (silent) states. Bottom: Spectrogram corresponding to the LFP showing high-frequency rhythms, mainly during the Up states. (C): Waveform average of the mean multiunit activity (MUA) signal aligned at the Down-to-Up state transition (dashed line at time 0). This average was used to calculate the maximum firing rate (red square; see Materials and Methods for details). Shade represents the SD. (DeE): Raw LFP recordings from right M1 at 7 and 20 months, respectively, in Wt (black) and 3xTg-AD (gray) mice. (FeG): Population data comparing the SO frequency between 3xTg-AD and Wt mice at 7 and 20 months of age, respectively, in the 4 cortical areas. (H): Population data showing the evolution of SO frequency between 7 and 20 months in both groups, with the 4 cortical areas pooled together. Bars and squares depict the mean, error bars are SE. * p < 0.0125 3xTg versus Wt mice and 7 versus 20 months of age in (H). Abbreviations: AD, Alzheimer’s disease; PrL, prelimbic cortex; M, months old; M1, primary motor cortex; S1, primary somatosensory cortex; V1, primary visual cortex. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 145 Table 1 Results of the repeated measures two-way ANOVA for each analyzed parameter in the 3xTg-AD mouse model Parameter Intrasubject factor: Intersubject factor: Intersubject factor: Interactions Cortical area group Age F p Effect F p Effect F p Effect Type F p Effect size (h2) size (h2) size (h2) size (h2) SO freq. 0.82 0.48 0.02 3.44 0.07 0.077 6.9 0.01 0.144 group  age 18.6 <0.0001 0.313 Up dur. 10.41 <0.0001 0.203 1.11 0.29 0.026 10.38 0.002 0.202 area  age 3.28 0.02 0.074 group  age 24.64 <0.0001 0.375 Down dur. 0.54 0.65 0.013 11.87 0.001 0.225 0.68 0.41 0.016 area  group 3.3 0.02 0.075 group  age 19.2 <0.0001 0.320 SO CV 2.82 0.04 0.064 9.01 0.005 0.180 2.63 0.11 0.06 group  age 10.28 0.003 0.201 FR Up 54.092 <0.0001 0.569 0.001 0.976 <0.0001 383.084 <0.0001 0.903 area  age 3.146 0.028 0.071 area  group 6.049 0.001 0.129 group  age 5.715 0.021 0.122 FR Down 29.533 <0.0001 0.425 2.585 0.116 0.061 350.961 <0.0001 0.898 area  group 5.077 0.002 0.113 FR rel. 49.981 <0.0001 0.549 6.696 0.013 0.140 34.921 <0.0001 0.460 area  age 3.563 0.016 0.080 area  group 3.013 0.033 0.068 group  age 13.072 0.001 0.242 Comparisons between groups and ages were performed with cortical area (PrL, prelimbic; M1, primary motor; S1, primary somatosensory; V1, primary visual) as the within- subjects variable, and age (7 and 20 mo of age) and group (controlewild typeeand 3xTg-AD) as between-subjects variables. Post hoc tests were performed with Student’s t-test for independent samples, corrected for multiple comparisons using the Bonferroni method (corrected a ¼ 0.0125). Bold text indicates p < 0.05. Key: SO, slow oscillation; freq, frequency; CV, coefficient of variation; FR, firing rate; rel, relative; dur, duration. Sanchez-Vives et al., 2010), several parameters of the oscillatory be higher in 3xTg mice (Fig. 1D, F and H), although this did not reach activity were quantified and compared between 3xTg-AD and the significance threshold established in the study. At 20 months of control mice: frequency of the SO, Up state duration, Down state age, however, the frequency of the SO was significantly lower in duration, the coefficient of variation of the Up-Down state cycle, 3xTg than in control mice (Fig. 1E, G, H). The reversal of the trend at the population firing rate (Fig. 1C) and the high-frequency content 20 months of age was due to the SO frequency being differently during the Up states. LFP recordings were obtained bilaterally, but modulated by age in each group, increasing in control animals be- no differences were found between hemispheres, neither for any tween 7 and 20 months of age, but remaining rather stabledwith a of the analyzed parameters and cortical areas nor in any of the 2 tendency to decreasedin 3xTg animals (Fig. 2H). groups; therefore, data of both hemispheres were averaged for The tendency of 7-month-old 3xTg mice to have a higher SO each cortical area and animal. frequency can be explained by the shorter Up states (Fig. 2D, F) because at this age, Down state duration did not differ between the 3.1. SO frequency and its variability are altered in the anesthetized 2 groups (Fig. 2A, C). Between 7 and 20 months of age, Down and Up 3xTg-AD mouse model state duration changed differentially in the 2 groups: Down state duration decreased in control animals but increased in the 3xTg The mean SO frequency differed between 3xTg and control ones (group  age interaction, Table 1; Fig. 2C), and a similar groups at 20 months, but not at 7 months. (group  age interaction, pattern was observed for the Up state duration, which decreased in Table 1, Fig. 1DeH). At 7 months, the frequency of the SO tended to control animals, but remained rather stabledwith a slight tendency Fig. 2. Up and Down state duration in the 3xTg-AD mouse. (A,B): Population data comparing Down state duration between 3xTg-AD and control mice at 7 and 20 months of age, respectively, in the 4 cortical areas. (D,E): Same as aforementioned, comparing Up state duration. (C,F): Population data showing the evolution of Down and Up state duration, respectively, between 7 and 20 months in both groups, with the 4 cortical areas pooled together. Bars and symbols depict the mean, error bars are SE. * p < 0.0125 3xTg versus control mice and 7 versus 20 months of age in (C and F). Abbreviations: AD, Alzheimer’s disease; PrL, prelimbic cortex; M, months old; M1, primary motor cortex; S1, primary somatosensory cortex; V1, primary visual cortex; Wt, wild type. 146 P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 to increasedin the 3xTg ones (group  age interaction, Table 1; Firing rate increased with age in both groups (Fig. 4). While the Fig. 2F). As a result of this differential evolution, at 20 months of age, mean firing rate during Down and Up states was comparable be- Down and Up state durations were significantly longer in 3xTg mice tween groups at 7 months of age (no main effect of group, Table 1; (Fig. 2B,C and E, F, respectively) whose SO frequency was conse- Fig. 4A, D), it had increased significantly by 20 months (main effect quently reduced (Fig. 1G, H). of age, Table 1; Fig. 4B, C), the mean firing rate during Up states Besides showing a lower SO frequency, the SO cycle was more bearing a greater increase in controls (group  age interaction, irregular in 20-month-old 3xTg mice than in their age-matched Table 1; Fig. 4E, F). As a consequence, between 7 and 20 months of controls (Fig. 3). This irregularity can be appreciated in raster age, the relative firing rate increased more markedly in control plots illustrating consecutive Up states aligned at their Down-to-Up animals, resulting in a significantly lower relative firing rate in 20- transition (Fig. 3A, B), which substantiate the elongation of Up and month-old 3xTg animals than in their age-matched controls, Down states in 20-month-old 3xTg mice and demonstrate a higher whereas at 7 months, such difference was not observed (group  temporal irregularity in this group. We corroborated these obser- age interaction, Table 1; Fig. 4GeI). vations at the population level: at 7 months of age, the mean CV of In both groups, the magnitude of the firing rate during the Down the SO did not differ between groups (Fig. 3C, E), but toward states was inversely related to its duration: longer Down states had 20 months of age, the CV of the SO evolved differently (group  age lower firing rates, in agreement with previous findings (Compte interaction, Table 1; Fig. 3E), augmenting in the 3xTg animals but et al., 2003; Sanchez-Vives et al., 2017) (Fig. 4J, K). However, remaining stable in the controls (Fig. 3E), leading to a higher CV of although the increase in firing rate during the Down states occur- the SO at 20 months of age in the 3xTg animals (Fig. 3D, E). ring between 7 and 20 months of age (Fig. 4C) happened in parallel to the shortening of Down states in control animals (see Fig. 3C), 3.2. Relative firing rate is reduced in the anesthetized 3xTg-AD this was not the case in the 3xTg animals, who showed an elon- mouse model at advanced ages gation of the Down states (see Fig. 3C). This suggests that the firing rate occurring during the Down states was less efficient in gener- The population firing rate during Up and Down states was ating Up states in the 3xTg mice at 20 months (Fig. 4A and Fig. 2A, calculated as the power change in the Fourier components between respectively). In addition, the firing rate during Down states and the 200 and 1500 Hz (Castano-Prat et al., 2017; Reig et al., 2010; Ruiz- SO frequency were related: higher firing rate during Down states Mejias et al., 2011), which is a good estimation of spiking activity of was associated with higher SO frequency (Fig. 4L, M). However, at the neurons surrounding the electrode tip (Mattia and Del Giudice, 20 months, there was a decrease in the correlation between the 2002). We also computed a relative measure of the firing ratedthe firing rate during the Down states and Down state duration in the relative firing ratedrepresenting the population spiking activity 3xTg animals from 7 to 20 months (Fig. 4J vs. Fig. 4K) along with a during Up states with respect to the spiking activity during the Down decline in the correlation between SO frequency and firing rate states (Fig. 1C; see Materials and Methods for further details). during the Down states in this group (Fig. 4L vs. Fig. 4M), illustrating While it is true Down states are considered relatively silent a disruption in the dynamical balance of the network. states with respect to Up states, there is some firing in Down states (Shu et al., 2003), and it has been proposed that this is critical for 3.3. Abnormal synchronization of beta and gamma frequencies the initiation of Up states (Compte et al., 2003). The firing rate during the Up states in the anesthetized 3xTg-AD mouse model during Down states is modulated by physiological conditions such as temperature (Reig et al. (2010) and that there is a relationship Finally, we studied the high-frequency content (beta and gamma between the firing rate in Down states and the frequency of the SO oscillations, 15e100 Hz) during Up and Down states in 3xTg and (D’Andola et al., 2018). control mice because synchronization in this band has been linked Fig. 3. Coefficient of variation (CV) of the SO in the 3xTg-AD mouse. (A,B): Representative raster plots showing aligned Up states (detected during 200 seconds) in M1 at 7 and 20 months of age, respectively, in one control (black) and one 3xTg (gray) mouse. (C,D): Population data comparing the CV of the SO between 3xTg-AD and control mice at 7 and 20 months of age, respectively, in the 4 cortical areas. (E): Population data showing the evolution of the CV of the SO between 7 and 20 months of age in both groups, with the 4 cortical areas pooled together. Bars and squares depict the mean, error bars are SE. * p < 0.0125 3xTg versus control mice and 7 versus 20 months of age in (E). Abbreviations: SO, slow oscillations; AD, Alzheimer’s disease; PrL, prelimbic cortex; M, months old; M1, primary motor cortex; S1, primary somatosensory cortex; V1, primary visual cortex; Wt, wild type. P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 147 Fig. 4. Population firing rate in the 3xTg-AD mouse. (A, B): Population data comparing firing rate (FR) during Down states between control (black) and 3xTg-AD mice (gray) at 7 and 20 months of age, respectively, in the 4 cortical areas. (D, E): Same as aforementioned, comparing FR during Up states. (G, H): Same as aforementioned, comparing the relative (Up/ Down) FR. (C): Population data showing the evolution of the FR during the Down states between 7 and 20 months of age, with both groups and the 4 cortical areas pooled together. (F, I): Population data showing the evolution of the FR during Up states and the relative FR, respectively, between 7 months and 20 months of age in both groups, with the 4 cortical areas pooled together. (J, K): Pearson’s correlation between the FR during Down states and the Down state duration at 7 and 20 months of age, respectively, in the 4 cortical areas. (L, M): Same as aforementioned, between the FR during Down states and the frequency of the SO. Bars and squares depict the mean, error bars are SE. * p < 0.0125 3xTg versus control mice and 7 versus 20 months of age in (C, F, I). Abbreviations: SO, slow oscillations; AD, Alzheimer’s disease; FR, firing rate; PrL, prelimbic cortex; M, months old; M1, primary motor cortex; S1, primary somatosensory cortex; V1, primary visual cortex; Wt, wild type; 7M, 7 months old; 20M, 20 months old. to cognitive and perceptual processes (Siegel et al., 2012; Wang, In both groups, PrL and M1 cortical areas presented a modula- 2010). We found that the mean power spectrum of 3xTg mice tion centered at 40e50 Hz during the Up states (Fig. 5AeD), significantly differed from that of controls at both ages, with het- although in 20-month-old 3xTg animals, the synchronization of the erogeneity depending on the cortical area. activity cortex did not form a peak in the power spectrum as clearly 148 P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 as in controls (especially in M1) (Fig. 5B, D). In S1 and V1 As control animals got older, the amplitude of the synchroni- (Fig. 5EeH), the synchronization peaks during the Up states were zation peak during the Up states was reduced in all cortical areas only clearly observed when calculating the excess power (see except for the S1 area, where its amplitude increased (Fig. 5B, D, F, Materials and Methods), revealing that these peaks occurred at and H, black arrowheads vs. black solid lines in the insets). We also lower frequencies than those in the PrL and M1 cortices, at both observed the decrease in the amplitude of the synchronization peak ages (Fig. 5EeH, vertical lines in the insets). Thus, anterior cortical with increasing age in the PrL and M1 cortices of 3xTg animals areas (PrL, M1) synchronized their activity during the Up states at (Fig. 5B, D, magenta arrowheads vs. red solid lines in the insets), but higher frequencies (ca. 50 Hz) than posterior areas (S1, V1) (ca. we detected group differences in S1 and V1: the large amplitude of 30e40 Hz), and this tendency was preserved in the 3xTg group. the S1 area synchronization peak observed in 20-month-old control Fig. 5. High frequency content of the SO in the 3xTg-AD mouse. (AeH): Mean raw power spectra of control (black) and 3xTg-AD (magenta) mice at 7 (left) and 20 (right) months of age in PrL (A, B), M1 (C, D), S1 (E, F), and V1 (G, H) cortices, during Up (solid lines) and Down states (dashed lines). Insets show excess power (ExcP) during the Up states, defined as the ratio between the mean power during Up states and the fit of the “1/f” decay of the power spectrum. Vertical lines in the insets indicate the power peak for each group. Arrowheads in the insets in (B, D, F, and H) indicate the frequency and amplitude of the power peak at 7 months. Shade is the SE. a.u. (arbitrary units). Abbreviations: SO, slow oscillations; AD, Alzheimer’s disease; PrL, prelimbic cortex; M1, primary motor cortex; S1, primary somatosensory cortex; V1, primary visual cortex; Wt, wild type. (For inter- pretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 149 animals was already present in 7-month-old 3xTg mice (Fig. 5E, F), (Funk et al., 2017; Zucca et al., 2017) and/or by the accumulation of suggesting it to be a sign of earlier aging; furthermore, V1 area activity-dependent potassium conductances (Compte et al., 2003; synchronization peak also increased with age in 3xTg animals Roopun et al., 2006; Sanchez-Vives and McCormick, 2000), mech- (Fig. 5H, arrowheads vs. solid lines in the insets). anisms that are activity-dependent (Sanchez-Vives et al., 2010). The The mean power spectrum differences between control and firing rates during Down and Up states increased in control and 3xTg animals during the Down states were not as pronounced and 3xTg animals as they got older, along with the expected shorter Up consistent across cortical areas (Fig. 5AeH, dashed lines) as during states in control animals, but not in 3xTg animals. This suggests that the Up states. At 7 months of age, the mean power spectrum during during the aging process, the physiological dynamic relationships Up states appeared to be displaced toward higher frequencies in regulating the oscillatory pattern were altered in the 3xTg-AD 3xTg animals with respect to controls (Fig. 5A, C, E, G, solid lines); mouse. Alterations in the oscillatory dynamics during slow wave this displacement was especially pronounced in the PrL and M1 sleep could be associated with sleep alterations in AD, associated cortices, but was apparent in all the cortical areas as a shift toward per se with cognitive decline (Ju et al., 2014). higher frequencies in the synchronization peak of the excess power An increase in cortical excitability associated with aging was also (Fig. 5A, C, E, G, vertical lines in the insets). However, at 20 months reflected in the firing rate during the Down states, which increased of age, the pattern was reversed, and the mean power spectrum of from 7 to 20 months of age in both groups of animals. This is 3xTg animals was then located at lower frequencies than that of coherent with the shortening of Down states that was observed in controls (Fig. 5B, D, F, H, solid lines). This inversion was due to the control animals as they got older, as high-frequency fluctuations fact that high frequency content during Up states evolved differ- during the Down state facilitate the generation of the subsequent ently in both groups as they got older: from 7 to 20 months, the Up state (Compte et al., 2003; Sanchez-Vives et al., 2017). However, synchronization peak of high frequencies during the Up states in the 3xTg-AD mouse model, the increase in firing rate during the moved toward higher frequencies in control animals, but toward Down states that occurs between 7 and 20 months of age was lower frequencies in 3xTg animals (Fig. 5B, D, F, H, arrowheads vs. accompanied by an elongation of the Down states, revealing an solid lines in the insets). increased threshold for recruiting the local network and trigger Up states. 4. Discussion Thus, the evolution of SO parameters as 3xTg-AD animals get older reveals difficulties to transition between states and thus in a We compared the Up and Down states of the spontaneously loss of flexibility of the cortical network. This could be the result of generated cortical SO, alongside the fast (beta and gamma) rhythms the cellular and synaptic alterations presented by 3xTg-AD animals during Up states between the 3xTg-AD mouse model of AD and at 20 months of age, such as the presence of extracellular Ab de- control mice at 7 and 20 months of age. 3xTg-AD mice presented posits (Oddo et al., 2003a; Sy et al., 2011) or the loss of dendritic alterations in the emergent cortical activity with respect to controls, spines (Bittner et al., 2010), and may be contributing to the cogni- and the electrophysiological phenotype of 3xTg-AD animals tive deficits reported in these animals, as higher cognitive functions changed with age in a different manner compared to that of control require a flexible modulation of the interactions between the animals. Almost all changes between 3xTg and control mice were neuronal groups belonging to a given cortical network (Bressler and detected at 20 months of age: 3xTg mice presented lower SO fre- Kelso, 2001; Fries, 2005). quency (consistent with longer Up and Down states), higher Up- Because the periodicity of the oscillatory cycle depends on the Down cycle variability, lower relative firing rate, and a shift of the interaction between the mechanisms involved in Up state initiation high-frequency rhythms toward lower frequencies, similarly to one and termination (reviewed by Neske, 2016), the SO variability in 20- of the electroencephalography hallmarks of AD in humans (Jeong, month-old 3xTg animals was increased. The loss of cholinergic re- 2004). The only change detected at 7 months of age was the ceptors in 3xTg-AD mice (Billings et al., 2005), also reported in reduction of Up state duration in 3xTg animals with respect to humans with AD (Coyle et al., 1983), may contribute as the rhyth- controls. The absence of more alterations in the cortical network micity of the SO diminishes when cortical cholinergic tone is physiology at that age is surprising, considering that 7-month-old reduced (Lorincz et al., 2015). 3xTg animals present with intracellular Ab immunoreactivity in the The synchronization peak of high (beta-gamma) frequencies neocortex (Oddo et al., 2003b) and behavioral deficits (Billings et al., during the Up state was detected at higher frequencies in 3xTg than 2005; Filali et al., 2012). We would suggest that larger alterations in control animals at 7 months old, but at lower frequencies at are needed to affect a pattern as robust and conserved as the SO 20 months. The dominant frequency of beta-gamma oscillations (Sanchez-Vives et al., 2017). These major alterations are indeed depends critically on the magnitude and temporal course of the present in 20-month-old 3xTg animals, in which Ab plaques and inhibitory feedback (Compte et al., 2008; Cunningham et al., 2004). neurofibrillary tangles are detectable throughout the neocortex The displacement of the high-frequency synchronization peak that (Oddo et al., 2003b; Sy et al., 2011), a considerable decrease in both groups presented as they got older suggests that although the dendritic spines density has occurred (Bittner et al., 2010), and the cortical network of 3xTg animals might be more excitable than that cerebral metabolism in cortical regions is already significantly of the control group at 7 months of age, at 20 months of age, this reduced (Nicholson et al., 2010). Altogether, this suggests that these network is more inhibited, and that while the excitability of the alterations are responsible for SO alterations detected in the pre- cortical network increases in the control animals with age, it de- sent work at 20 months, and that the intracellular Ab accumulation creases in the case of the 3xTg animals. Reduced cortical excitability is not enough to cause significant changes in the SO parameters in the 20-month-old 3xTg animals is consistent with the lower characterized here. relative firing rate and with the longer Up states due to decreased The increase in the slow oscillation frequency that occurred activity-dependent potassium conductances (Sanchez-Vives et al., between 7 and 20 months of age in control animals did not take 2010). In fact, lower high frequencies are commonly described place in the 3xTg animals. In the 3xTg-AD mouse, the Up state EEG spectral alterations in patients with AD (Dierks et al., 2000; duration remained stable throughout the 2 studied ages, whereas Huang et al., 2000; Lizio et al., 2011) and have been related to the the Down state duration significantly increased between 7 and deterioration of global cognitive function (Jeong, 2004; Kowalski 20 months of age. The transition from the Up state to the Down et al., 2001; Sloan et al., 1995). The shift toward lower fre- state has been proposed to be mediated by inhibitory interneurons quencies of the beta-gamma synchronization peak in the 20- 150 P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 month-old 3xTg animals is reminiscent of this alteration and in- This work was supported by Ministerio de Economía y Com- dicates that their electrophysiological phenotype is consistent petividad (Spain, BFU2017-85048-R), FLAG-ERA (PCIN-2015-162- with that of humans with AD and other mouse models of the C02-01), and CERCA Programme/Generalitat de Catalunya. disease (Castano-Prat et al., 2017). This suggests that the alteration of the inhibitory component of the network described in advanced Appendix A. Supplementary data AD stages in humans (Francis et al., 1999) could be relevant to the pathophysiology of the disease. Supplementary data associated with this article can be found, in It has been reported that ketamine anesthesia enhances gamma the online version, at https://doi.org/10.1016/j.neurobiolaging.2019. power (Hong et al., 2009; Hunt et al., 2010; Lazarewicz et al., 2009). 02.009. This could be a confounding factor in the reported results; however, both Wt and 3xTg-AD animals were recorded under the same anesthesia, and therefore, we should attribute the observed dif- References ferences to changes in the network. Still, we should mention that a potential interaction between the anesthesia and the disease could Akay, M., Wang, K., Akay, Y.M., Dragomir, A., Wu, J., 2009. Nonlinear dynamical eventually occur, although we did not observe differences in the analysis of carbachol induced hippocampal oscillations in mice. Acta Pharmacol. Sin. 30, 859e867. required anesthesia doses. This is one of the problems that are Billings, L.M., Oddo, S., Green, K.N., McGaugh, J.L., LaFerla, F.M., 2005. Intraneuronal encountered when comparing brain activity in Wt versus diseased Abeta causes the onset of early Alzheimer’s disease-related cognitive deficits in animals: not only diseased animals could respond differently to the transgenic mice. Neuron 45, 675e688. Bittner, T., Fuhrmann, M., Burgold, S., Ochs, S.M., Hoffmann, N., Mitteregger, G., anesthesia, but if studying wakefulness, there could be also a Kretzschmar, H., LaFerla, F.M., Herms, J., 2010. Multiple events lead to dendritic different alert state. In previous studies in Dyrk1A (Down Syndrome spine loss in triple transgenic Alzheimer’s disease mice. PLoS One 5, e15477. model) in which a decreased gamma generation was observed Bressler, S.L., Kelso, J.A.S., 2001. Cortical coordination dynamics and cognition. Trends Cogn. Sci. 5, 26e36. during Up states, the same was the case when awake mice were Castano-Prat, P., Perez-Zabalza, M., Perez-Mendez, L., Escorihuela, R.M., Sanchez- recorded (Ruiz-Mejias et al., 2016). Still, further studies with Vives, M.V., 2017. Slow and fast neocortical oscillations in the senescence- chronically implanted animals will be useful in the case of AD to accelerated mouse model SAMP8. Front. Aging Neurosci. 9, 141. Compte, A., Sanchez-Vives, M.V., McCormick, D.A., Wang, X.J., 2003. Cellular and establish whether the alterations that we describe in high fre- network mechanisms of slow oscillatory activity (<1 Hz) and wave propaga- quencies remain in the awake state. tions in a cortical network model. J. Neurophysiol. 89, 2707e2725. In conclusion, SO parameters evolved in a different manner in Compte, A., Reig, R., Descalzo, V.F., Harvey, M.A., Puccini, G.D., Sanchez-Vives, M.V., 2008. Spontaneous high-frequency (10-80 Hz) oscillations during up states in control and 3xTg-AD animals although they were quite similar at the cerebral cortex in vitro. J. Neurosci. 28, 13828e13844. 7 months of age. In particular, 20-month-old 3xTg animals pre- Coyle, J.T., Price, D.L., DeLong, M.R., 1983. Alzheimer’s disease: a disorder of cortical sented with lower SO frequency, longer Up and Down states, higher cholinergic innervation. Science 219, 1184e1190. SO variability, lower relative firing rate, and a shift toward lower Cunningham, M.O., Whittington, M.A., Bibbig, A., Roopun, A., LeBeau, F.E.N., Vogt, A., Monyer, H., Buhl, E.H., Traub, R.D., 2004. A role for fast rhythmic bursting frequencies of the mean power spectrum during the Up states with neurons in cortical gamma oscillations in vitro. Proc. Natl. Acad. Sci. U. S. A. 101, respect to age-matched Wt controls, the latter of which mimics one 7152e7157. of the features in humans with AD, corroborating the suitability of Destexhe, A., Hughes, S.W., Rudolph, M., Crunelli, V., 2007. Are corticothalamic “up” states fragments of wakefulness? Trends Neurosci. 30, 334e342. the 3xTg mouse as a model for this disorder. The evolution of the SO Dierks, T., Jelic, V., Pascual-Marqui, R.D., Wahlund, L., Julin, P., Linden, D.E., parameters as 3xTg animals get older also points to an alteration of Maurer, K., Winblad, B., Nordberg, A., 2000. Spatial pattern of cerebral glucose the network mechanisms involved in the transitions between Up metabolism (PET) correlates with localization of intracerebral EEG-generators in Alzheimer’s disease. Clin. Neurophysiol. 111, 1817e1824. and Down states, resulting in a less flexible cortical network that D’Andola, M., Weinert, J.F., Mattia, M., Sanchez-Vives, M.V., 2018. Modulation of struggles to transition between states. Decreased excitability and slow and fast oscillations by direct current stimulation in the cerebral cortex flexibility of the cortical network in older 3xTg-AD animals is the in vitro. bioRxiv 246819. Filali, M., Lalonde, R., Theriault, P., Julien, C., Calon, F., Planel, E., 2012. Cognitive and result of cellular and synaptic alterations and may be contributing non-cognitive behaviors in the triple transgenic mouse model of Alzheimer’s to the cognitive deficits reported in this model. disease expressing mutated APP, PS1, and Mapt (3xTg-AD). Behav. Brain Res. Our results help to unveil the cortical network functional alter- 234, 334e342. Francis, P.T., Palmer, A.M., Snape, M., Wilcock, G.K., 1999. The cholinergic hypothesis ations in AD, which probably underlie also cognitive deficits of Alzheimer’s disease: a review of progress. J. Neurol. Neurosurg. Psychiatry 66, commonly observed in AD, and may have practical implications in 137e147. preclinical research, such as allowing the tracking of the impact of Fries, P., 2005. A mechanism for cognitive dynamics: neuronal communication novel molecules aimed at modifying the course of neuronal dys- through neuronal coherence. Trends Cogn. Sci. 9, 474e480. Funk, C.M., Peelman, K., Bellesi, M., Marshall, W., Cirelli, C., Tononi, G., 2017. Role of functions in this AD model. In other pathologies such as Fragile X somatostatin-positive cortical interneurons in the generation of sleep slow syndrome (Castano-Prat et al., under review), we have observed waves. J. Neurosci. 37, 9132e9148. how genetic and pharmacological interventions reverse slow wave Gimenez-Llort, L., Blazquez, G., Canete, T., Johansson, B., Oddo, S., Tobena, A., LaFerla, F.M., Fernandez-Teruel, A., 2007. Modeling behavioral and neuronal alterations along with other manifestations of the disease. Because symptoms of Alzheimer’s disease in mice: a role for intraneuronal amyloid. we have found that SO alterations increase with age, along with Neurosci. Biobehav. Rev. 31, 125e147. other structural and functional changes described in other works, Hong, L.E., Summerfelt, A., Buchanan, R.W., O’Donnell, P., Thaker, G.K., Weiler, M.A., Lahti, A.C., 2009. Gamma and delta neural oscillations and association with specific changes in cortical oscillatory networks may be helpful to clinical symptoms under subanesthetic ketamine [online]. Neuro- predict to what extent cognitive dysfunctions are likely to evolve psychopharmacology 35, 632. with the disease. Huang, C., Wahlund, L., Dierks, T., Julin, P., Winblad, B., Jelic, V., 2000. Discrimination of Alzheimer’s disease and mild cognitive impairment by equivalent EEG sources: a cross-sectional and longitudinal study. Clin. Neurophysiol. 111, Disclosure 1961e1967. Hunt, M.J., Falinska, M., qe˛ ski, S., Wójcik, D.K., Kasicki, S., 2010. Differential effects produced by ketamine on oscillatory activity recorded in the rat hippocampus, The authors declare no conflict of interest. dorsal striatum and nucleus accumbens. J. Psychopharmacol. 25, 808e821. Jeong, J., 2004. EEG dynamics in patients with Alzheimer’s disease. Clin. Neuro- Acknowledgements physiol. 115, 1490e1505. Ju, Y.-E.S., Lucey, B.P., Holtzman, D.M., 2014. Sleep and Alzheimer disease pathology– a bidirectional relationship. Nat. Rev. Neurol. 10, 115e119. The authors thank Dr. C. Gonzalez-Liencres for editing Katzner, S., Nauhaus, I., Benucci, A., Bonin, V., Ringach, D.L., Carandini, M., 2009. assistance. Local origin of field potentials in visual cortex. Neuron 61, 35e41. P. Castano-Prat et al. / Neurobiology of Aging 79 (2019) 142e151 151 Kowalski, J.W., Gawel, M., Pfeffer, A., Barcikowska, M., 2001. The diagnostic value of Ruiz-Mejias, M., Ciria-Suarez, L., Mattia, M., Sanchez-Vives, M.V., 2011. Slow and fast EEG in Alzheimer disease: correlation with the severity of mental impairment. rhythms generated in the cerebral cortex of the anesthetized mouse. J. Clin. Neurophysiol. 18, 570e575. J. Neurophysiol. 106, 2910e2921. LaFerla, F.M., Oddo, S., 2005. Alzheimer’s disease: abeta, tau and synaptic Ruiz-Mejias, M., Martinez de Lagran, M., Mattia, M., Castano-Prat, P., Perez- dysfunction. Trends Mol. Med. 11, 170e176. Mendez, L., Ciria-Suarez, L., Gener, T., Sancristobal, B., Garcia-Ojalvo, J., Lazarewicz, M.T., Ehrlichman, R.S., Maxwell, C.R., Gandal, M.J., Finkel, L.H., Siegel, S.J., Gruart, A., Delgado-Garcia, J.M., Sanchez-Vives, M.V., Dierssen, M., 2016. Over- 2009. Ketamine modulates theta and gamma oscillations. J. Cogn. Neurosci. 22, expression of Dyrk1A, a Down syndrome candidate, decreases excitability and 1452e1464. impairs gamma oscillations in the prefrontal cortex. J. Neurosci. 36, 3648e3659. Lizio, R., Vecchio, F., Frisoni, G.B., Ferri, R., Rodriguez, G., Babiloni, C., 2011. Elec- Sanchez-Vives, M.V., Mattia, M., 2014. Slow wave activity as the default mode of the troencephalographic rhythms in Alzheimer’s disease. Int. J. Alzheimers Dis. cerebral cortex. Arch. Ital. Biol. 152, 147e155. 2011, 927573. Sanchez-Vives, M.V., McCormick, D.A., 2000. Cellular and network mechanisms of Lorincz, M.L., Gunner, D., Bao, Y., Connelly, W.M., Isaac, J.T.R., Hughes, S.W., rhythmic recurrent activity in neocortex. Nat. Neurosci. 3, 1027e1034. Crunelli, V., 2015. A distinct class of slow (w0.2-2 Hz) intrinsically bursting layer Sanchez-Vives, M.V., Mattia, M., Compte, A., Perez-Zabalza, M., Winograd, M., 5 pyramidal neurons determines UP/DOWN state dynamics in the neocortex. Descalzo, V.F., Reig, R., 2010. Inhibitory modulation of cortical up states. J. Neurosci. 35, 5442e5458. J. Neurophysiol. 104, 1314e1324. Mattia, M., Del Giudice, P., 2002. Population dynamics of interacting spiking neu- Sanchez-Vives, M.V., Massimini, M., Mattia, M., 2017. Shaping the default activity rons. Phys. Rev. E. Stat. Nonlin. Soft Matter Phys. 66, 51917. pattern of the cortical network. Neuron 94, 993e1001. Mukovski, M., Chauvette, S., Timofeev, I., Volgushev, M., 2007. Detection of active Shu, Y., Hasenstaub, A., McCormick, D.A., 2003. Turning on and off recurrent and silent states in neocortical neurons from the field potential signal during balanced cortical activity. Nature 423, 288e293. slow-wave sleep. Cereb. Cortex 17, 400e414. Siegel, M., Donner, T.H., Engel, A.K., 2012. Spectral fingerprints of large-scale Neske, G.T., 2016. The slow oscillation in cortical and thalamic networks: mecha- neuronal interactions. Nat. Rev. Neurosci. 13, 121e134. nisms and functions. Front. Neural Circuits 9, 88. Sloan, E.P., Fenton, G.W., Kennedy, N.S., MacLennan, J.M., 1995. Electroencephalog- Nicholson, R.M., Kusne, Y., Nowak, L.A., LaFerla, F.M., Reiman, E.M., Valla, J., 2010. raphy and single photon emission computed tomography in dementia: a Regional cerebral glucose uptake in the 3xTG model of Alzheimer’s disease comparative study. Psychol. Med. 25, 631e638. highlights common regional vulnerability across AD mouse models. Brain Res. Squirrell, D., 2015. An in Vivo Electrophysiological and Computational Analysis of 1347, 179e185. Hippocampal Synaptic Changes in the Alzheimer’s Disease Mouse. The Uni- Oddo, S., Caccamo, A., Kitazawa, M., Tseng, B.P., LaFerla, F.M., 2003a. Amyloid versity of Manchester, Manchester, UK. deposition precedes tangle formation in a triple transgenic model of Alz- Steriade, M., Timofeev, I., 2003. Neuronal plasticity in thalamocortical networks heimer’s disease. Neurobiol. Aging 24, 1063e1070. during sleep and waking oscillations. Neuron 37, 563e576. Oddo, S., Caccamo, A., Shepherd, J.D., Murphy, M.P., Golde, T.E., Kayed, R., Steriade, M., Nunez, A., Amzica, F., 1993. A novel slow (< 1 Hz) oscillation of Metherate, R., Mattson, M.P., Akbari, Y., LaFerla, F.M., 2003b. Triple-transgenic neocortical neurons in vivo: depolarizing and hyperpolarizing components. model of Alzheimer’s disease with plaques and tangles: intracellular Abeta and J. Neurosci. 13, 3252e3265. synaptic dysfunction. Neuron 39, 409e421. Steriade, M., Amzica, F., Contreras, D., 1996. Synchronization of fast (30-40 Hz) Paxinos, G., Franklin, K., 2008. The Mouse Brain in Stereotaxic Coordinates, third ed. spontaneous cortical rhythms during brain activation. J. Neurosci. 16, 392e417. Academic Press, Cambridge, MA. Sy, M., Kitazawa, M., LaFerla, F.M., 2011. The 3xTg-AD mouse model: reproducing Querfurth, H.W., LaFerla, F.M., 2010. Alzheimer’s disease. N. Engl. J. Med. 362, and modulating plaque and tangle pathology. In: Peter Paul De Deyn, D.V.D. 329e344. (Ed.), Animal Models of Dementia. Springer, pp. 469e482. Reig, R., Mattia, M., Compte, A., Belmonte, C., Sanchez-Vives, M.V., 2010. Tempera- Wang, X.J., 2010. Neurophysiological and computational principles of cortical ture modulation of slow and fast cortical rhythms. J. Neurophysiol. 103, rhythms in cognition. Physiol. Rev. 90, 1195e1268. 1253e1261. Wong, P.C., Cai, H., Borchelt, D.R., Price, D.L., 2002. Genetically engineered mouse Roopun, A.K., Middleton, S.J., Cunningham, M.O., LeBeau, F.E., Bibbig, A., models of neurodegenerative diseases. Nat. Neurosci. 5, 633e639. Whittington, M.A., Traub, R.D., 2006. A beta2-frequency (20-30 Hz) oscillation Zucca, S., D’Urso, G., Pasquale, V., Vecchia, D., Pica, G., Bovetti, S., Moretti, C., in nonsynaptic networks of somatosensory cortex. Proc. Natl. Acad. Sci. U. S. A. Varani, S., Molano-Mazón, M., Chiappalone, M., Panzeri, S., Fellin, T., 2017. An 103, 15646e15650. inhibitory gate for state transition in cortex. Elife 6.