Age-related dysregulation of homeostatic control in neuronal microcircuits

Neuronal homeostasis prevents hyperactivity and hypoactivity. Age-related hyperactivity suggests homeostasis may be dysregulated in later life. However, plasticity mechanisms preventing age-related hyperactivity and their efficacy in later life are unclear. We identify the adult cortical plasticity response to elevated activity driven by sensory overstimulation, then test how plasticity changes with age. We use in vivo two-photon imaging of calcium-mediated cellular/synaptic activity, electrophysiology and c-Fos-activity tagging to show control of neuronal activity is dysregulated in the visual cortex in late adulthood. Specifically, in young adult cortex, mGluR5-dependent population-wide excitatory synaptic weakening and inhibitory synaptogenesis reduce cortical activity following overstimulation. In later life, these mechanisms are downregulated, so that overstimulation results in synaptic strengthening and elevated activity. We also find overstimulation disrupts cognition in older but not younger animals. We propose that specific plasticity mechanisms fail in later life dysregulating neuronal microcircuit homeostasis and that the age-related response to overstimulation can impact cognitive performance.

Neuronal homeostasis prevents hyperactivity and hypoactivity.Age-related hyperactivity suggests homeostasis may be dysregulated in later life.However, plasticity mechanisms preventing age-related hyperactivity and their efficacy in later life are unclear.We identify the adult cortical plasticity response to elevated activity driven by sensory overstimulation, then test how plasticity changes with age.We use in vivo two-photon imaging of calcium-mediated cellular/synaptic activity, electrophysiology and c-Fos-activity tagging to show control of neuronal activity is dysregulated in the visual cortex in late adulthood.Specifically, in young adult cortex, mGluR5-dependent population-wide excitatory synaptic weakening and inhibitory synaptogenesis reduce cortical activity following overstimulation.In later life, these mechanisms are downregulated, so that overstimulation results in synaptic strengthening and elevated activity.We also find overstimulation disrupts cognition in older but not younger animals.We propose that specific plasticity mechanisms fail in later life dysregulating neuronal microcircuit homeostasis and that the age-related response to overstimulation can impact cognitive performance.
Homeostatic regulation of neuronal activity promotes stable neural circuit function by preventing extreme spiking 1 .Regulation of neuronal activity occurs through homeostatic plasticity mechanisms 2,3 operating at the network, cell and synapse [1][2][3][4] .However, the efficacy of such mechanisms in later life is unclear [5][6][7] .Regulation of activity in the aged brain is important, as neural hyperactivity may increase susceptibility to neurodegeneration 8,9 and impair cognition 10 .Age-related hyperactivity occurs in multiple species and brain regions, suggesting that regulatory plasticity mechanisms preventing hyperactivity in young adult cortex may fail in later life 8,[10][11][12][13][14] .Understanding such early age-related changes may facilitate targeting of vulnerable processes.
Current understanding of plasticity mechanisms preventing neural hyperactivity in the adult cortex is limited.The majority of in vivo work investigating mature regulatory plasticity processes has focused on adaptation to reduced activity [1][2][3]15 . Howver, work in neuronal culture finds plasticity processes can modify the synaptic excitatory-to-inhibitory (E:I) ratio to compensate for prolonged chemically induced hyperactivity 16,17 .The E:I ratio can be modified through processes such as synaptic downscaling [17][18][19][20] .This cell-autonomous mechanism broadly reduces synaptic strength and could prevent saturation of synaptic weights and maintain neuronal firing within a stable dynamic range 21 .These population-level changes in synaptic strength can be described by a single multiplicative or additive scaling factor 17,22 .Population-level weakening of excitatory synapses can involve group I metabotropic glutamate receptors (mGluR1 and mGluR5) and the mGluR1/mGluR5-Homer1 signaling pathway in juvenile and adult cortex [23][24][25] .However, the expression/efficacy of such regulatory plasticity during aging remains unclear.
We next measured more fine-scale activity changes in L2/3 neurons by recording after each hour-long overstimulation session in anesthetized mice.We again quantified the AUC of ΔF/F 0 calcium signal per second (Supplementary Fig. 1c) and, consistent with c-Fos measures (Supplementary Fig. 1a,b), found an initial elevation in total activity across S1-3 of day 1 (Supplementary Figs.1c and 2f).The acute elevation was followed by a reduction to values less than baseline by S4 of day 1 (Supplementary Figs.1c and 2f).More prolonged overstimulation (days 2-3) resulted in total activity levels being less than baseline after each day of overstimulation (Supplementary Figs.1c and 2f).The overstimulation response was driven by changes in ΔF/F 0 amplitude, frequency or both (Supplementary Fig. 2f,h,j,l,n and Supplementary Table 8) and was evident in both visually evoked and spontaneous activity (Supplementary Fig. 2f,h,j,l,n and Supplementary Table 8).
Experience-dependent changes in neuronal activity can occur independently of N-methyl-d-aspartate receptors (NMDARs) 22 .To test if the overstimulation response was NMDAR dependent, we used 3-(2-carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP), an NMDAR antagonist blocking some forms of Hebbian plasticity for up to 24 h without affecting visual function [31][32][33] .We administered CPP 24 h before and immediately before overstimulation, and then we performed imaging (Methods and Supplementary Fig. 1d,e).The acute increase (S3 of day 1) and subsequent decrease in activity were NMDAR independent (Supplementary Fig. 1e).Our results suggest overstimulation drives an NMDAR-independent adaptation of neuronal activity in young adult V1.

Synaptic plasticity after overstimulation in young adult cortex
Changes in neuronal activity may occur through shifts in the synaptic E:I ratio 2 .We measured the E:I ratio with electrophysiological recordings of miniature excitatory postsynaptic potentials (mEPSPs) and miniature inhibitory postsynaptic potentials (mIPSPs) from L2/3 excitatory neurons in brain slices containing V1 in overstimulated and control mice (Fig. 1d-f and Supplementary Fig. 1f-i).Mice received 3 d of overstimulation, with tissue collected 24 h later so that plasticity, rather than the immediate impact of overstimulation, could be assessed (Supplementary Fig. 1f).The E:I ratio was estimated using the AUC of mEPSPs/ mIPSPs (Methods).Overstimulation reduced the E:I ratio (Fig. 1d) due to increased mIPSP frequency (Fig. 1e and Supplementary Fig. 1g) and reduced mEPSP amplitude (Fig. 1f).Increased synaptic inhibition and decreased synaptic excitation were also evident in spontaneous EPSPs/ IPSPs (sEPSPs/sIPSPs) (Supplementary Fig. 1j,m).Overstimulation did not modify mIPSP amplitude, sIPSP amplitude (Supplementary Fig. 1h, k), mEPSP frequency, sEPSP frequency (Supplementary Fig. 1i,l), action potential rheobase or membrane properties (Supplementary Table 7).Thus, overstimulation reduced the E:I ratio in young adult V1 through specific synaptic mechanisms.
To test for structural correlates of functional synaptic changes, we first measured dendritic spine size as a proxy for synaptic strength 3 .We tested whether overstimulation impacted spine size in L2/3 V1 neurons in Thy1-eGFP mice 34 (Methods and Fig. 1g).To assay inhibitory circuitry, we used immunofluorescence approaches and labeled vesicular γ-aminobutyric acid transporter (VGAT) at Thy1-eGFP-positive neurons (Methods and Fig. 1g).Overstimulation decreased spine size, and this was blocked by the mGluR5 inhibitor MTEP 23,25 (Fig. 1h).Reduced spine size was accompanied by increased density of VGAT puncta at L2/3 Thy1-eGFP + neurons of overstimulated mice (Fig. 1i).VGAT and spine density were not modified by MTEP treatment (Fig. 1i and Supplementary Fig. 1n).Our results suggest that mGluR5-dependent synaptic weakening occurs following overstimulation.
Plastic changes in excitatory synaptic strength may work with other regulatory mechanisms, such as changes in synaptic inhibition or modifications of intrinsic excitability, to modulate activity levels following elevated activity 20 .Whether different processes are more susceptible to aging remains unclear 5 , as do the implications for network function and cognition 26 .Age-related downregulation of one plasticity mechanism may be compensated by another 27 .For example, some animal models express an age-related shift from homeostatic to more Hebbian-like plasticity 28 .Thus, rather than a general decline, plasticity may be differentially regulated across mechanisms, pathways, cell types and/or brain regions during aging 5 .
We found age-related changes in the regulation of neuronal activity between young (3 months) and late (12 months) adulthood in mouse primary visual cortex (V1).This age range tested plasticity differences before widespread neuronal hyperactivity and pronounced sensorimotor deficits 5 .We used visual overstimulation to drive elevated cortical activity and combined two-photon (2-P) calcium imaging with c-Fos-activity tagging and electrophysiology, to identify plasticity mechanisms in young adult cortex and investigate age-related changes.We found age-related dysregulation of population-level multiplicative excitatory synaptic weakening and formation of inhibitory inputs onto excitatory neurons.In young adult cortex, these mechanisms reduced synaptic E:I ratios and dampened cortical activity following overstimulation.Neuronal and subcellular calcium imaging found overstimulation reduced dendritic spine activity and strengthened functional associations between excitatory and inhibitory neuronal assemblies in young adult cortex.Older animals exhibited failures in identified plasticity mechanisms, showing greater synaptic strengthening after overstimulation, which increased the E:I ratio and strengthened associations between excitatory neurons.Overstimulation also disrupted subsequent cognitive performance in older but not younger animals.A positive allosteric modulator (PAM) of mGluR5 signaling prevented the negative impact of overstimulation on cognition in older animals, while blockade of mGluR5-dependent processes promoted the negative impact of overstimulation on cognition in young adult animals.Our results find specific synaptic plasticity mechanisms fail in later life leading to dysregulation of neuronal activity homeostasis.We propose that disrupted homeostatic plasticity has consequences for network plasticity and that the age-related response to sensory overstimulation can impact cognitive performance.

Neuronal overstimulation response in young adult cortex
Age-related hyperactivity may be due to declining regulatory plasticity mechanisms 1,8,[10][11][12][13][14]29,30 . To cature emerging changes and age-related plasticity differences, we focused on ages preceding the classical definition of rodent old age, widespread neuronal hyperactivity and pronounced sensorimotor deficits 5 .We therefore tested for plasticity mechanisms triggered by elevated cortical activity in young adult mice (3 months old) and the efficacy of plasticity in later life (8-12 months old).
We developed a visual overstimulation paradigm to elevate activity of layer 2/3 (L2/3) excitatory neurons in V1, a locus of adult plasticity 2,3 (Fig. 1a,b and Supplementary Fig. 1a,b).Overstimulation involved repeated 1-h sessions (S) with a 40-Hz light flicker in the home cage (four times per day, 1 h on/1 h off) across the light cycle (Methods and Fig. 1a).We found that two 1-h sessions of overstimulation acutely elevated cortical c-Fos activity (Methods and Supplementary Fig. 1a,b).We then performed longitudinal in vivo 2-P recordings of calcium-mediated activity at L2/3 excitatory neurons expressing GCaMP6s in anesthetized mice (Methods and Fig. 1a,b).We imaged before (day 0) and after each overstimulation day (days 1-3) (Methods and Fig. 1a-c).Recordings were made during light-flicker (matching overstimulation stimulus) and resting-state activity (Methods and Fig. 1a,b).Neuronal activity was estimated using the area under the curve (AUC) of the calcium We next asked if weakened excitatory synaptic strength is captured by multiplicative scaling factors, similarly to reports during downward firing rate homeostasis 22 .We used a standard approach 3,35 to test if multiplicative values could scale control distributions down to measures made after overstimulation (Methods and Supplementary Fig. 1o-y).For mEPSP amplitude (Supplementary Fig. 1o,p) and spine size (Supplementary Fig. 1q,r), control values could be scaled to the overstimulation distribution.We next tested if overstimulation modified expression of the AMPA receptor subunit GluA2, the synaptic scaffolding molecules GRIP1 and HOMER1 and the Ca 2+ /calmodulin-dependent kinase eEF2K, as these synaptic molecules have been linked to both population-level multiplicative synaptic plasticity 3,36,37 and long-term depression (LTD) 38,39 .Immunofluorescence experiments 3 measured expression levels per unit of spine size of GluA2 (Supplementary Fig. 1s,t), GRIP1 (Supplementary Fig. 1u,v), HOMER1 (Supplementary Fig. 1w,x) and eEF2k (Supplementary Fig. 1y) at Thy1-eGFP + dendrites in L2/3 of V1  , 'C') for all conditions.Data were obtained from 7 animals for c, 19 animals for d-f and 9 animals for panels g-i.*P < 0.05, ***P < 0.001; NS, not significant.Two-sided tests were used throughout.Detailed statistics are reported in Supplementary Table 1.Data are presented as the mean ± s.e.m. with individual data points (gray dots) or as a cumulative distribution.

Age-related dysregulation of response to overstimulation
We next asked if the neuronal activity response to overstimulation (Fig. 1c and Supplementary Figs.1c and 2f) changed with age by repeating our in vivo calcium imaging experiments in adult mice aged 3-12 months (Methods and Fig. 2a-c).We found an age-related decline in neurons expressing reduced activity after overstimulation (Methods, Fig. 2d,e and Supplementary Fig. 2a).In 3-month-old mice, ~70% of neurons had reduced activity after overstimulation (Fig. 2d and Supplementary Fig. 2a).This percentage declined with age, reaching values similar to those of controls by 12 months (~45%; Fig. 2d,e and Supplementary Fig. 2a).At the population level, animals aged 3-8 months showed reduced activity after day 1 of overstimulation, while 12-month-old animals did not (Fig. 2f).In older animals (aged 8-12 months), activity increased across further overstimulation sessions (Fig. 2f); as a result, in 12-month-old animals, activity was greater than control values on day 3 of overstimulation (Fig. 2f and Supplementary Fig. 2b).In contrast, a consistent reduction in activity was maintained across overstimulation sessions in 3-month-old animals (Fig. 2f and Supplementary Fig. 2b).Activity levels in age-matched anesthetic control animals were stable during imaging (Fig. 2g,h, Supplementary Fig. 2b and Supplementary Table 8).However, in agreement with previous work 8,29 , the percentage of neurons with high-frequency calcium-mediated neuronal activity increased with age (Methods and Supplementary Fig. 2c).
We investigated more fine-scale age-dependent activity changes recording from active neurons after each overstimulation session (Fig. 2g-k and Supplementary Fig. 2d-s).Both young animals (3 months) and animals in later adulthood (12 months) showed an acute (S1-3, day 1) elevation in total activity (Fig. 2g-i) that was not evident in anesthetic controls (Fig. 2g,h).The acute elevation was evident, and similar, for both age groups, for measures of visually evoked (Fig. 2j) and spontaneous activity (Fig. 2k) but occurred slightly earlier in older animals (Fig. 2h-k and Supplementary Fig. 2f-o).After overstimulation, twelve-month-old animals had total activity values greater than baseline (Fig. 2h-k and Supplementary Fig. 2g), with similar changes in visually evoked (Fig. 2j and Supplementary Fig. 2i,k) and spontaneous (Fig. 2k and Supplementary Fig. 2m,o) activity.Increased activity was driven by changes in ΔF/F 0 amplitude, frequency or both (Supplementary Fig. 2g,i,k,m,o and Supplementary Table 8).Age-related differences after overstimulation may be due to slower rates of adaptation in older animals.However, this was not the case, as animals in later adulthood continued to have elevated activity levels even after extended (6 d) periods of overstimulation (Methods, Fig. 2h-k and Supplementary Fig. 2g,i,k,m,o).Activity levels in young adults returned to baseline values following prolonged overstimulation (Fig. 2g,i,j,k and Supplementary Fig. 2f,h,j,l,n).In late adults, activity levels remained elevated 3 d after cessation of overstimulation but did decline relative to measures over days 3-6 (Supplementary Fig. 2p,q).Overstimulation did not modify orientation (Supplementary Fig. 2r) or direction (Supplementary Fig. 2s) selectivity in young or late adult animals, suggesting basic features of sensory encoding are unaffected despite elevated activity.Our results find the adaptive response to overstimulation becomes dysregulated between early and late adulthood (Fig. 2a-k and Supplementary Fig. 2f-o), across a period when more neurons exhibit high-frequency activity (Supplementary Fig. 2c).Furthermore, the overstimulation response in animals in later adulthood is tenacious and not due to slower adaptive plasticity.

Age-related dysregulation of overstimulation-driven synaptic plasticity
We next tested whether changes in the plasticity mechanisms identified in young adult animals accounted for the age-related response to overstimulation.We performed electrophysiological recordings of mEPSPs/ mIPSPs from L2/3 excitatory neurons in control and overstimulated mice across our age range and calculated the E:I ratio (Methods and Fig. 3a).Older animals (8-12 months) exhibited a different synaptic plasticity response to overstimulation than young adults (3 months; Fig. 3a).By 12 months, excitatory neurons exhibited an increased E:I ratio after overstimulation relative to controls, rather than the decreased E:I ratio measured in young adults (Fig. 3a).Age impacted the inhibitory and excitatory synaptic plasticity response to overstimulation (Fig. 3a-c and Supplementary Fig. 3a-d).In 12-month-old mice, mIPSP frequency was no longer increased after overstimulation, and instead showed a decrease relative to controls (Fig. 3b and Supplementary Fig. 3b).This effect occurred in concert with a progressive age-related increase in mEPSP amplitude after overstimulation, to values greater than control levels in 8-month-old and 12-month-old overstimulated mice (Fig. 3c and Supplementary Fig. 3c,d).Similar changes in synaptic excitation and inhibition were found after overstimulation in measures of spontaneous EPSPs/IPSPs from 12-month-old animals (Supplementary Fig. 3g,h).We saw no age-related effects on mEPSP frequency (Supplementary Fig. 3e), mIPSP amplitude (Supplementary Fig. 3f), rheobase or passive membrane properties (Supplementary Table 9) after overstimulation, suggesting age impacts specific synaptic mechanisms.
Overstimulation drives elevated activity (Fig. 2a-k) and an increased synaptic E:I ratio in older animals (Fig. 3a-c).We tested for structural synaptic correlates at neurons recently active in vivo (Methods and Fig. 3d).The immediate early gene cFos can approximate in vivo neuronal activity and allow recently active neurons to be fluorescently tagged 2,40 .We therefore adopted two c-Fos-based strategies.First, an in vivo viral expression approach that labeled the entire neuron, and second, an ex vivo staining method to validate our viral strategy.The virally expressed C-FOS-eYFP construct labeled the dendritic and axonal arborization of neurons expressing c-Fos (Methods and Fig. 3d) and reported similar c-Fos expression to immunofluorescence measures (Supplementary Fig. 3i).We prepared brain slices from animals expressing C-FOS-eYFP and measured VGAT puncta and dendritic spines at c-Fos-positive neurons in 3-month-old, 8-month-old and 12-month-old control and overstimulated animals (Fig. 3d-h and Supplementary Figs.1f and 3j-w).We observed increased VGAT density with age in control animals, suggesting inhibitory synaptogenesis persists into late adulthood (Supplementary Fig. 3j).In 3-month-old mice, overstimulation resulted in increased VGAT density at c-Fos-positive neurons (Fig. 3e and Supplementary Fig. 3k).The inhibitory response to overstimulation declined with age (Fig. 3e and Supplementary Fig. 3k-m), so that 12-month-old overstimulated mice exhibited a small but significant reduction in VGAT density relative to controls (Fig. 3e and Supplementary Fig. 3m).In 3-month-old overstimulated animals, increased VGAT density was evident at neurons expressing both low and higher levels of c-Fos (Supplementary Fig. 3n-p and Methods).However, in overstimulated late adult animals, reduced VGAT density was specific to neurons with higher levels of c-Fos expression (Supplementary Fig. 3q), suggesting that loss of VGAT puncta in older animals may contribute to the elevated neuronal activity following overstimulation (Supplementary Fig. 3q and Fig. 2a-k).We next measured dendritic spines at c-Fos-positive branches (Fig. 3f and Supplementary Fig. 3r-t), and found a progressive age-related increase in spine size after overstimulation (Fig. 3f and Supplementary Fig. 3r-t),   without a change in density (Supplementary Fig. 3u).Spine density and size were stable across our chosen age range in control animals (Supplementary Fig. 3v-w).Increased excitatory synaptic strength without a balanced change in inhibition may drive the heightened neuronal activity we measured in vivo in late adult animals after overstimulation (Fig. 2a-k).We tested the correlation between synaptic plasticity (structural E:I ratio; Methods) and estimates of neuronal activity (using c-Fos) on a cell-by-cell basis.We found an age-related increase in c-Fos intensity after overstimulation positively correlated with the structural synaptic E:I ratio (Fig. 3h).Our results suggest that in 12-month-old animals, overstimulation increases the E:I ratio, as opposed to the decrease seen in young adult animals (Fig. 3a).The increased E:I ratio is due to reduced inhibitory synaptic plasticity and a shift from excitatory synaptic weakening to strengthening after overstimulation (Fig. 3i).

Age-related dysregulation of in vivo dendritic spine plasticity
We investigated the functional identity of synapses driving the age-related overstimulation response from population-level weakening to synaptic strengthening.We used in vivo 2-P imaging to measure    4a-e).We collected a baseline session (Fig. 4d,e) and then imaged after a day of overstimulation (four sessions, post-stimulation; Methods and Fig. 4d,e).We extracted spine responses using published methods 41,42 (Supplementary Fig. 4a-c) and compared activity before and after overstimulation (Fig. 4d-f).
We measured the population response of calcium signals at spines (Fig. 4f and Supplementary Fig. 4d-e) during light-flicker stimuli similar to the home-cage overstimulation.In 3-month-old animals, overstimulation decreased population-level spine activity in vivo (Fig. 4f and Supplementary Fig. 4d).In contrast, overstimulation increased spine activity in 12-month-old animals (Fig. 4f and Supplementary Fig. 4e).These results mirrored electrophysiological and spine size measures (Figs.1f,h and 3c,f) and suggest the overstimulation response shifts from population-level excitatory synaptic weakening to strengthening with age in vivo.
To better understand how functional spine plasticity relates to neuronal activity in vivo, we compared the spine plasticity response to changes in activity at the adjacent dendrite following overstimulation.Dendritic activity measured in this way is thought to approximate global cellular activity 3 .Similar dendritic activity changes were accompanied by markedly different changes in spine activity after overstimulation in young and late adult animals (Fig. 4g).In young animals, spine activity was reduced at branches with elevated activity after overstimulation (Fig. 4g).In contrast, the activity of spines in older animals was positively correlated with changes in the activity of dendritic branches (Fig. 4g).These results suggest that similar changes in dendritic activity are accompanied by synaptic weakening in young adults and strengthening in older animals.We tested whether the shift from synaptic weakening to strengthening occurred at all spines, or those most sensitive to overstimulation (Fig. 4h).We measured spine responsivity during the overstimulation stimulus and compared to the number of events during spontaneous activity 2,43 (Methods, Fig. 4h and Supplementary Fig. 4f).We defined visually responsive spines as those with time-locked responses to >20% of visual stimulus trials and classified remaining spines as nonvisual (that is, non-responsive to the overstimulation stimulus; Supplementary Fig. 4f).Both visual and nonvisual spines were present on all imaged branches and did not exhibit clustering (Supplementary Fig. 4g,h).

Identified mechanisms account for age-related overstimulation response
We next modeled the overstimulation response using plastic excitatory and inhibitory mechanisms and feedforward excitatory input (Methods and Supplementary Fig. 4j).The plasticity of synaptic weights was governed by Hebbian learning and modulated by inhibition and homeostatic adaptation, which scaled excitatory synaptic weights (Supplementary Fig. 4j).This model was compared to a late adulthood simulation where excitatory homeostasis and inhibitory plasticity were reduced during the overstimulation phase (Methods and Supplementary Fig. 4j).
In the young adult simulation, overstimulation and Hebbian plasticity strengthened weights between neurons with a coincident increase of presynaptic and postsynaptic activity and weakened weights between neurons with uncorrelated activity.The impact of increased inhibition and excitatory synaptic homeostasis gave weight changes (Supplementary Fig. 4k) aligned with experimental observations (Fig. 4i).Reducing excitatory synaptic homeostasis and the plasticity of inhibition resulted in weight changes akin to those measured in older animals (Fig. 4i,j and Supplementary Fig. 4k,l).Specifically, our results show increased synaptic strengthening in the most visually responsive inputs and reduced synaptic weakening in less visually responsive inputs (Supplementary Fig. 4k,l).
We investigated the contribution of excitatory and inhibitory plasticity by changing each process independently (Supplementary Fig. 4m-r).Starting conditions matched our late adulthood simulation (Supplementary Fig. 4m,p), and then inhibitory (Supplementary Fig. 4m-o,s) and excitatory (Supplementary Fig. 4p-r,t) plasticity mechanisms were progressively reinstated, and the impact of overstimulation on synaptic weights was assessed.Increasing inhibitory plasticity, during reduced excitatory homeostasis, reduced the strengthening of all weights, but with most prominent changes at more visually responsive weights (Supplementary Fig. 4m-o,s).Increasing levels of excitatory synaptic homeostasis, during reduced inhibitory plasticity, led to a weakening of all weights relative to the late adulthood simulation (Supplementary Fig. 4p-r), with most prominent changes at the least visually responsive inputs (Supplementary Fig. 4t).These results were explained by Hebbian plasticity having most impact on the highly visual synapses, which have highest covariance with postsynaptic activity (Methods and equation ( 3)).As a result, increasing inhibitory plasticity most strongly suppresses the potentiation of visual synapses.In contrast, the homeostatic term weakens all synapses, but weakening of nonvisual synapses is more pronounced, because Hebbian potentiation at these synapses is small.In summary, both plasticity processes are required to capture features of our experimental findings.
We investigated whether inhibitory neuronal activity following overstimulation was modified by age.Overstimulation did not modify the inhibitory population activity in young adults (Fig. 5c).However, inhibitory activity increased after overstimulation in older animals (Fig. 5c), but this did not dampen the activity of excitatory neurons (Supplementary Fig. 5d).Our earlier experiments found reduced VGAT density at excitatory neurons in older animals after overstimulation (Fig. 3e).We wondered whether a reduction in VGAT may occur at inhibitory neurons in older animals after overstimulation, possibly facilitating elevated inhibitory activity (Fig. 5c).We measured VGAT puncta at GAD-positive neurons in young and late adult animals (Methods and Supplementary Fig. 5e-h) and found the size (Supplementary Fig. 5e), but not density (Supplementary Fig. 5f), of VGAT puncta was reduced in late adult animals after overstimulation.In young adult animals, the size and density of VGAT puncta were similar to those at control levels in GAD-positive neurons after overstimulation (Supplementary Fig. 5g,h).Our results suggest that overstimulation may lead to weakening of synaptic inhibition onto inhibitory neurons in late adult animals, possibly contributing to the elevated inhibitory activity in vivo.
We investigated the age-related network-level response to overstimulation at assemblies of excitatory and inhibitory neurons (Fig. 5d-i), and estimated functional associations via pairwise calcium correlations (Methods and Fig. 5d,e) 2 .We considered neurons correlated if coefficients were significant and positive 2,42,43 .For each excitatory neuron, we calculated average pairwise correlation strength with other excitatory or inhibitory neurons in the local cortical region during visual stimulation (Methods and Fig. 5e).In young adult animals, overstimulation increased association strength between excitatory and inhibitory neurons but had little impact on association strength between only excitatory cells (Fig. 5f,g).In contrast, overstimulation increased association strength between excitatory neurons in late adult animals, but not between excitatory and inhibitory cells (Fig. 5h,i).Overstimulation did not impact associations between inhibitory neurons in either young or late adult animals (Supplementary Fig. 5i).Our results suggest that age modifies the network-level plasticity response of functional assemblies to overstimulation, so that rather than developing stronger associations with inhibitory cells, excitatory neurons in older animals become more strongly associated with each other.

Overstimulation drives age-related disruption of cognition
We next tested if overstimulation impacted cognitive performance at different ages (3-12 months) using a touchscreen task involving visual discrimination (rodent continuous performance task (rCPT)) 45 (Fig. 6a and Supplementary Fig. 6a).In this task, animals make a correct nose poke to a rewarded visual stimulus (Hit), withhold responses to unrewarded stimuli (correct rejection (CR)), fail to respond (miss) or respond to an unrewarded stimulus (mistake).These behaviors generate the hit rate (HR) (HR = Hits/(Hits + Misses)), false alarm rate (FAR) (FAR = Mistakes/(Mistakes + Correct rejections) ) and performance (HR − FAR) metrics.The slope of a linear fit to performance scores across sessions gave a learning rate (Methods, Fig. 6a and Supplementary Fig. 6b).
All ages (3-12 months) learnt the task (Fig. 6a,c, Supplementary Fig. 6a,b and Table 6) and thus we tested for an interaction between overstimulation and cognitive performance.Animals received the overstimulation paradigm and were tested after a rest period (~1 h; Methods and Fig. 6b).Overstimulation did not impact rCPT performance in 3-month-old mice (Fig. 6c).However, as animals aged, overstimulation resulted in progressively worse performance and elevated c-Fos (Fig. 6c,d).The results could not be explained by age-related sensorimotor deficits, reward association or motivation (Supplementary Fig. 6c-e).
Young adult plasticity mechanisms may confer resilience to overstimulation by preventing elevated neuronal activity, which can impair rodent cognition 10 .In young cortex, we found excitatory synaptic weakening after overstimulation is mGluR5 dependent (Fig. 1h).We therefore administered the noncompetitive mGluR5 inhibitor MTEP to young adult mice, to block this form of excitatory synaptic weakening, as well as other mGluR5-dependent processes 46 (Fig. 1h) and repeated the overstimulation and behavioral experiments (Methods and Fig. 6e).MTEP alone did not impact performance (Fig. 6e); however, MTEP plus overstimulation reduced performance of young adult animals (Fig. 6e).All animals showed similar reward latency (Supplementary Fig. 6f), correct choice latency (Supplementary Fig. 6g) and trial completion (Supplementary Fig. 6h).However, although MTEP alone did not modify these variables, MTEP plus overstimulation led to a small increase in the reward collection latency (Supplementary Fig. 6f) and decreased trial number (Supplementary Fig. 6h), suggesting a possible reduction in motivation.
We next asked if positive allosteric modulation of mGluR5 could reduce the negative impact of overstimulation on cognitive   48 , which stabilizes mGluR5 at the plasma membrane 47,48 .Treatment with VU0409551 alone did not impact rCPT performance in unstimulated older animals (Fig. 6f).However, the negative impact of overstimulation on cognitive performance in older animals was reduced if animals received VU0409551 (Fig. 6f).This suggests that certain PAMs of mGluR5 signaling may protect against the negative impact of overstimulation on cognition in older animals (Fig. 6f).

Discussion
We find synaptic plasticity mechanisms triggered by sensory overstimulation in young adults are dysregulated in late adulthood.In young adult cortex, overstimulation results in excitatory synaptic weakening and elevated synaptic inhibition, reducing the synaptic E:I ratio and dampening neuronal activity.This is accompanied by weakening of dendritic spine activity and strengthened associations between excitatory and inhibitory neurons after overstimulation in vivo.In contrast, overstimulation led to excitatory synaptic strengthening, modified network plasticity and a failure to recruit increased synaptic inhibition in late adult animals.Overstimulation also disrupted cognitive performance in older but not younger animals.The negative impact of overstimulation on cognition could be activated by blocking mGluR5-dependent processes in young adult animals and alleviated by a PAM of mGluR5 signaling in older animals.We find specific plasticity mechanisms are associated with age-related dysregulation of neuronal microcircuit homeostasis, and that the age-related response to sensory overstimulation impacts cognitive performance.

Age-related impairments in plasticity processes
Aging may cause a general decline in plasticity processes; alternatively, specific forms of plasticity may be more vulnerable to aging 5 .We asked whether specific synaptic plasticity processes, triggered by elevated neuronal activity, exhibit an age-related decline, focusing on early-to-late adulthood time points to capture onset of mechanistic changes in plasticity 5 .We find age-related plasticity dysregulation can be unmasked by overstimulation.Age-related deficits occurred in two specific processes, namely, population-level excitatory synaptic weakening and increased synaptic inhibition onto excitatory neurons.Dysregulation of these mechanisms meant that repeated overstimulation drove elevated neuronal activity in older animals.Our work is supported by others reporting age-related neuronal hyperactivity can impair neural circuit function 49 , have negative consequences for cognition 10 and occur across brain regions and species, including hum ans 8,[10][11][12][13][14]29,30,50 .

Alignment with theories of plasticity
Age-related failures in regulatory plasticity may have consequences for other forms of plasticity, or trigger compensatory processes 51 .In addition, multiple forms of homeostatic and Hebbian plasticity may be driven by overstimulation 23,52,53 .Our modeling found age-related  6.Data are presented as the median and interquartile range (in the box plots, the box indicates quartiles, and whiskers represent the 10th and 90th percentiles), with individual data points (gray dots).

Article
https://doi.org/10.1038/s41593-023-01451-zplasticity effects were consistent with dysregulated homeostasis, leading to unconstrained Hebbian synaptic strengthening.However, such synaptic strengthening in older animals may occur due to an age-related shift in the synaptic potentiation threshold 52 .This scenario aligns with metaplasticity theories 52 .In fact, work in younger animals (postnatal day (P) 21-35) found excitatory synaptic weakening after sensory manipulation follows metaplasticity rules 23,52 .However, we found activity changes in young adult animals did not involve classical NMDAR-dependent plasticity mechanisms.Furthermore, metaplasticity theories do not fully capture the inhibitory synaptic changes we measured in older animals, although there are few studies in this area.The strengthening of spines in older animals may involve an independent plasticity process.Alternatively, the population-wide weakening we measured may be a global effect acting against the synaptic strengthening that would occur in its absence.Future work may consider how multiple plasticity mechanisms are triggered, both in adulthood and later senescence.For example, one study on sensory deprivation in juvenile mice (P24-30) found NMDAR-dependent and NMDAR-independent plasticity in response to the level of activity manipulation 53 .Previous work describes homeostatic adaptation during sensory-evoked activity, typically following sensory deprivation [54][55][56][57] .Few studies have investigated the response to sensory-driven increases in activity, and those that have typically induce deprivation first 22,23 .We find sensory-evoked and spontaneous activity show an acute increase, then downward adaptation in young adults, with both phases independent of NMDARs.Overstimulation also reduced the E:I ratio in young adult animals, via mGluR5-dependent weakening of population-level excitatory synaptic strength and increased synaptic inhibition.Functional, structural and molecular measures of excitatory synaptic weakening occurred at the population level and were captured by multiplicative scaling rules, consistent with previous work 3,22,35 .Population-level excitatory synaptic weakening is dependent on dephosphorylation of the GluA2 AMPA receptor subunit, via Homer1a-mGluR1/mGluR5-mediated signaling cascades 18,23,25 .We found reduced GluA2, GRIP1 and HOMER1 expression in young adults following overstimulation.A key molecular feature of mGluR5-dependent LTD (eEf2k expression) was unaffected by overstimulation in young adult animals 39 .Together, the molecular, synaptic and cellular changes are consistent with homeostatic adaptation 20,21,25 .However, we cannot exclude that other non-NMDAR, mGluR5-dependent processes may weaken excitatory synapses and work in concert with increased synaptic inhibition.

Possible molecular mechanisms
In older animals, overstimulation led to increased neuronal activity and a reduced inhibitory plasticity response.Brain-derived neurotrophic factor 58 and nitric oxide signaling 59 are known to increase inhibitory synaptic drive after elevated excitatory spiking.Brain-derived neurotrophic factor signaling is driven by activity-dependent transcription factors, such as Npas4, which coordinate E:I balance and decline with age 60,61 .The Npas4-Homer1a signaling pathway mediates homeostatic plasticity mechanisms 62 and, in aged mice, elements of this pathway decrease 63 .Recruitment of synaptic inhibition may also be modified by age-related changes in synaptic adhesion proteins, such as neuroligins, which recruit inhibitory synapses onto excitatory neurons 64 and show alterations during aging 65 .Future work may investigate if the inhibitory response is affected by age-related alterations to such molecular processes.

Time course of plasticity processes
The plasticity time course likely depends on age, manipulation and studied system.In juvenile animals, homeostatic plasticity involving cellular or synaptic changes occurs 24-72 h after deprivation [66][67][68] .Elevated activity, following eyelid closure then reopening, also triggers plasticity across 24-48 h (ref.22).In adults, retinal lesions drive upscaling within 24 h, with in vivo recovery of activity evident 18 h after deprivation 69 .Such approaches also trigger upscaling-like plasticity at inhibitory neurons 3 by 6 h (ref.70).Adult monocular enucleation drives spine size increases (8-24 h) preceding recovery of activity (48 h) 2,3 .Molecular changes underpinning scaling can occur within 4 h of pharmacological challenge in culture systems 71 .We find relatively fast functional spine changes, occurring after overstimulation (~10 h).Overstimulation may drive plasticity at a different rate to deprivation paradigms.For example, whisker overstimulation (24 h) drives rapid increases in inhibitory synapses 72 .In addition, juvenile mice (P21-35) subjected to dark exposure followed by light reexposure showed weakened synaptic strength within 2 h 23 .Future work may test plasticity timescales across manipulations of differing severity 53 .

Overstimulation and cognition
Age-related impairments in Hebbian plasticity can reduce cognitive performance 73 .How dysregulated homeostasis impacts cognition is less clear 26 .Overstimulation impaired cognitive performance in older, but not younger, animals.Eight-month-old animals showed a loss of E:I plasticity, without elevated E:I ratios, but still exhibited reduced performance after overstimulation.This, in combination with measures in 12-month-old animals, suggests age dysregulates plasticity.Short bouts of visual stimulation (1 h per day at 40 Hz) can improve cognition in Alzheimer's disease models 74 .We used prolonged and repeated overstimulation to drive a strong plasticity response.Thus, one possibility is that the duration of overstimulation has different effects on plasticity and, ultimately, cognition.For example, a shorter repeated stimulus may drive Hebbian-like potentiation and/or neuroprotective processes 74,75 , while prolonged stimulation could upregulate homeostatic plasticity.
Age-related neuronal hyperactivity correlates with impaired rodent cognition 10 .Antiepileptic drugs improve cognition in aged rodents and humans with mild cognitive impairment 76 .We found blockade of mGluR5-dependent processes rendered young adult animals susceptible to the negative impact of overstimulation on cognition, while a PAM of mGluR5 reduced the negative impact of overstimulation on cognition in older animals.Thus, mGluR5-dependent mechanisms may modulate the interaction between overstimulation and cognition.However, further work is required to determine if mGluR5 is directly involved in the age-related alteration to overstimulation.We propose that age-related impairments in regulatory homeostatic plasticity processes, when challenged with sensory overstimulation, promote elevated cortical activity and may subsequently impair cognition.

Animals
Experiments were conducted according to the United Kingdom Animals (Scientific Procedures) Act 1986 and approved by the UK Home Office.We used male and female adult (P60-360) C57BL/6J mice, or adult (P60-100) Thy1-eGFP mice on a C57BL/6J background ( JAX 007788, The Jackson Laboratory; Fig. 1g-i and Supplementary Fig. 1n,q-y).Ages were defined as: young adult (3 months), mature adult (8 months) and late adult (12 months) 5 .Mice were matched for sex and age within experimental groups, but randomly assigned to conditions, and housed (temperature, 21 ± 2 °C; humidity, 55% ± 10%) with littermates on a 12 -h light-dark cycle.Longitudinal imaging data were matched for time within that animal's light cycle.All mice had access to water and food ad libitum, except during the rCPT ('Behavioral experiments').For animal numbers in each experiment, see figure legends and Supplementary Tables 1-12.

Visual overstimulation
LED strips around the home cage emitted warm white light (~2,200 K) at 40 Hz and a 50% duty cycle from the start of the light cycle for four bouts of 1 h on, alternated with 1 h off, over 8 h.Stimulation covered 3-6 d (Results).Controls received matched handling and sham stimulation.To test acute impact of overstimulation, AAV1-C-FOS-eYFP (Vigene Biosciences, 1.67 × 10 13 genome copies (GC) per ml) was injected into V1, and after recovery, in vivo 2-P imaging ('Functional and structural imaging') measured c-Fos expression after two 1 h bouts of overstimulation (Supplementary Fig. 1a,b).

Functional and structural imaging
Calcium imaging used a custom 2-P microscope (INSS) with a resonance scanner (Cambridge Technology) and a high-power objective Z-piezo stage (Physik Instrumente), using a Chameleon Vison II laser (Coherent) and a water immersion ×16, 0.8 NA, 3.0 mm WD objective (Nikon).Data were acquired with an 800 MHz digitizer (National Instruments) and pre-processed with a custom-programmed field programmable gate array (National Instruments).Mesoscopic imaging localized the visual cortex 2 .Average laser power was <50 mW and acquisition used ScanImage (Vidrio Technologies).Acquisition parameters for excitatory neurons were: 512 × 512 pixels (340 × 340 µm; Figs. 1 and 2 and Supplementary Figs. 1 and 2), spanning six sections with a 40-µm z-step at 5-7 Hz (Figs. 1 and 2).Spine imaging involved a single time series across 512 × 512 pixels (130 × 130 µm; Fig. 4) at 30 Hz, as did excitatory and inhibitory cell recordings (Fig. 5 and Supplementary Fig. 5).The excitation wavelength was set to 920 nm to excite GCaMP6s.At an excitation wavelength of 920 nm, cyRFP (only expressed in inhibitory neurons) is also excited.Emitted light from cyRFP was filtered by an ET605/70M band-pass filter.After functional imaging, the excitation wavelength was set to 1,080 nm to excite mCherry in excitatory neurons 41 .For imaging of C-FOS-eYFP, excitation wavelength was 920 nm (Supplementary Fig. 1a,b).For the in vivo imaging, mice were briefly anesthetized with oxygen and isoflurane (1-1.5%)mix and head-fixed.For all 2-P calcium imaging experiments, mice were anesthetized (0.5% isoflurane) and anesthetic level was the same for all age groups.Mice habituated to the setup before imaging sessions.
A coarse timeline determined time points of interest (Figs.1c and  2d-f).A more fine-scale time course then investigated within sessions and across days (Fig. 2g-k and Supplementary Fig. 2).For the coarse timeline, data were acquired during a baseline before overstimulation (day 0) and after each day of overstimulation or control: 24 h (day 1), 48 h (day 2) and 72 h (day 3; Figs.1a-c and 2a-f).For the fine-scale time-course analysis, data were collected during baseline (day 0, S1-4) and after each 1 h overstimulation session (four times per day for 3 d), so that S4 (Fig. 2g-k and Supplementary Fig. 2f-o) was equivalent to the coarse imaging (Figs.1c and 2f).Data were also collected across S1-4 on day 6 (Fig. 2g-k and Supplementary Fig. 2f-o).In a subset of 12-month-old mice, data were collected 3 d after overstimulation (Supplementary Fig. 2p,q).For spine imaging, data were acquired during baseline (Fig. 4), and the next day, 2 h after the 8-h-long overstimulation protocol (Fig. 4).For functional assemblies (Fig. 5), mice were imaged at baseline (Fig. 5), and then after 2 d of overstimulation (Fig. 5).Recordings consisted of 6-12 trials including three to six trials of 85 s of LED stimulation at 40 Hz for 3 s on, interspersed by 2 s off.This was alternated with three to six 85 s of recording in the dark.Measures of orientation and direction selectivity indices (OSI and DSI) were collected and analyzed in response to randomly presented drifting-grating stimuli 2,43 .

Calcium imaging analysis
For cellular and spine analysis, full-frame data were registered (Moco, ImageJ v1.53t) (ref.77) and analyzed by at least three experimenters who were blind to condition and time course.Cellular (and background) regions of interest (ROIs) were drawn.Cells with filled nuclei due to GCaMP overexpression were excluded.Signal processing removed slow changes in fluorescence, normalized to the background and thresholded (15%).For excitatory and inhibitory neurons, activity was estimated as the AUC of the ΔF/F 0 trace per second 2,42,43,78 .Measures of ΔF/F 0 amplitude and frequency were made using peak detection in MATLAB 3 .Total cellular activity was taken as activity recorded during LED stimulation and in the dark.In a subset of analyses, activity was normalized to baseline levels, by either subtracting the AUC of each cell's activity from its baseline (Figs.1c and 2f), or by normalizing the AUC of active cells and spines to the mean of the baseline (Figs.2g,h, 4 and 5 and Supplementary Fig. 2).Deconvolution of calcium traces used standard approaches 79 .Analysis estimating highly active neurons (Supplementary Fig. 2c) was adapted from published approaches 8 .Calcium transient frequency was calculated and the percentage of neurons falling into the low (minimum-Q1), https://doi.org/10.1038/s41593-023-01451-zwith weak c-Fos expression to VGAT density at dendrites with greater c-Fos levels (Supplementary Fig. 3n-q).We normalized c-Fos values from control animals to the mean of each distribution within age group and pooled across ages (Supplementary Fig. 3n).We considered c-Fos levels less than or equal to the 30th percentile of the control distribution to be weakly expressing c-Fos (Supplementary Fig. 3n) and compared VGAT density values at these dendrites with greater levels (>30th percentile) of c-Fos expression (Supplementary Fig. 3o-q).

Computational modeling
We simulated a two-compartment somatic and dendritic model (Supplementary Fig. 4j).Neuronal activity depends on somatic input (I soma ) and excitatory and inhibitory dendritic inputs (I dend ) according to equation (1): Here, τ is the time constant of network integration (τ = 10), and [] + denotes rectification.Each input component is given, respectively, as per equation ( 2): I ffw is the feedforward input to the neuron (I ffw = 0.5) and r exc and r inh are the aggregate rate of excitatory and inhibitory inputs impinging on the somatic compartment, which are kept in balance (r exc = r inh ).N is the number of excitatory inputs from presynaptic sources, while r i and w i denote the activity and the connection weight of the i-th source, respectively.Total inhibitory input at this component, I inh , is modeled as a single aggregate input, summarizing the combined effect of the activity and weight of presynaptic inhibitory sources.Overstimulation changes the activity of the network, due to primarily feedforward mechanisms that depend on the visual responsiveness of the neuron.The initial firing rate of the i-th neuron in the network is given by r i = r 0 + r m , where r 0 is the baseline firing rate before overstimulation (r 0 = 1), and r m is the modulation of activity by visual stimulation.r m alternates between on and off states, with the level of modulation depending on the visual responsiveness of the neuron (v): r m = v for T on and r m = 0 for T off .T on = T off = 100ms.Visual responsiveness for each neuron, v i , is drawn randomly from a uniform distribution between 0 (nonvisual) and 1 (the most visual).The postsynaptic neuron receives input from presynaptic neurons with levels of visual responsiveness.The initial weight of connections is set as: w i = 1 + ξ, where ξ is an independent and identically distributed random variable drawn from a normal distribution with N(0, 0.1).
Evolution of the activity of the neuron in equation ( 1) depends on the plasticity of its weights (w in equation ( 2)).The weight of the dendritic input from the i-th presynaptic excitatory neuron, w i , is evolving, in turn, according to equation (3): where r is the postsynaptic activity described in equation ( 1) and r 0 is the baseline activity.The first plasticity term describes the Hebbian component of learning, which comprises presynaptic rates (r i ) and postsynaptic activity (r).The strength of this learning is controlled by the parameter A 1 , which is set to A 1 = 1 × 10 −4 for both young and old animals and is modulated by inhibition (1-I inh ).In both young and old animals, at a baseline state I young inh = I old inh = 0.2.The second term on the right-hand side in equation ( 3) is a homeostasis term, which downscales the weights as a function of the postsynaptic activity of the neuron, and its strength is controlled by the downscaling parameter A 2 .
Following the experimental results, we find that in young adult animals, both inhibitory synaptic mechanisms and excitatory synaptic weakening may regulate the network activity after overstimulation (Figs.2f and 3a-c).Therefore, after overstimulation, inhibition is upregulated in the model to a higher value by allowing I young inh = 0.4 , while the downscaling parameter is set to A young 2 = 0.56 × 10 −4 .Experimental findings show that in old animals excitatory synaptic weakening and the increased inhibitory response to sensory overstimulation are reduced (Figs.3a-c and 4f).This is modeled by a lack of increased inhibition (fixed), with I old inh remaining at the same baseline value of 0.2, and a reduction in levels of downscaling in old animals, where where v i is the visual responsivity of the respective presynaptic neuron, as described above, and ζ(t) is an independent and identically distributed random variable drawn from N(0, 0.1).α is the feedback parameter, which changes the activity of presynaptic neurons as a function of the postsynaptic activity.We chose α = 0.1.This models how the change in the activity of postsynaptic neurons is reflected in the activity of presynaptic neurons, which are connected in a recurrent network.The activity of the postsynaptic neuron in turn would change as a result.Excitatory and inhibitory inputs to the somatic compartment are modeled as r exc (t) = r 0 + ζ(t) and r inh (t) = r 0 + ζ(t), where ζ(t) is drawn from a uniform distribution of (0, 0.1).
To test the contribution of plasticity mechanisms (Supplementary Fig. 4m-r), initial conditions matched our late adult simulation (Supplementary Fig. 4m,p).We then progressively reinstated plasticity mechanisms individually and measured the impact on synaptic weights of simulated overstimulation (Supplementary Fig. 4m-r).Specifically, we increased the parameter controlling the strength of dendritic inhibition (I inh ) from its late adult value (0.2) to 0.5 (corresponding to strong inhibition in young adult): I inh = [0.2, 0.3, 0.4, 0.5] (Supplementary Fig. 4m-o).We then kept dendritic inhibition fixed at late adult values and modulated the downscaling homeostatic term ( A 2 ).The strength of downscaling was increased from its weak value in the late adult to stronger values: [0.6, 0.7, 0.8, 0.9] (Supplementary Fig. 4p-r).For these simulations, we quantified the normalized change in the plasticity of the synaptic weights following modulation of either inhibition (Supplementary Fig. 4s) or downscaling (Supplementary Fig. 4t) as a function of the late adulthood simulation.

Behavioral experiments
Behavioral tests used Bussey-Saksida touchscreen operant chambers (Campden Instruments).ABET II (Campden Instruments) and WhiskerServer (Cambridge University Technical Services) software were used to run the boxes.Behavioral testing was conducted as previously described 45 .Three days before testing, animals were food restricted to 85-90% of free feeding weight.This weight was maintained throughout testing and water was available ad libitum 45 .Following habituation to the setup, mice underwent three stages of the rCPT (Fig. 6a and Supplementary Fig. 6a) 45 .Each session lasted 45 min a day, or 100 rewards.Stage 1 and 2 criteria required mice to earn 100 rewards within the 45-min session.For stage 1, mice were trained to touch a white-outlined square on the screen.For stage 2, the white frame was replaced by a target (S+) stimulus (horizontal or vertical line) presented for 5 s.For stage 3, a non-target (S−) stimulus (a 'snowflake') was introduced and randomly presented on 50% of the trials.
For stages 2-3, entry into the reward magazine initiated a delay period of 2 s before proceeding to the next intertrial interval (3-5 s).

A old 2 = 0 . 48 × 10 − 4 .
Each neuron receives input from N = 200 presynaptic excitatory sources.The evolution of weights and the postsynaptic response is modeled for T = 10, 000, and simulations are performed with time steps of dt = 1.The response of excitatory presynaptic sources projecting to dendrites evolves in time according to equation (4):

Table 2
. Data are presented as the mean ± s.e.m.