Cellular and synaptic phenotypes lead to disrupted information processing in Fmr1-KO mouse layer 4 barrel cortex

Sensory hypersensitivity is a common and debilitating feature of neurodevelopmental disorders such as Fragile X Syndrome (FXS). How developmental changes in neuronal function culminate in network dysfunction that underlies sensory hypersensitivities is unknown. By systematically studying cellular and synaptic properties of layer 4 neurons combined with cellular and network simulations, we explored how the array of phenotypes in Fmr1-knockout (KO) mice produce circuit pathology during development. We show that many of the cellular and synaptic pathologies in Fmr1-KO mice are antagonistic, mitigating circuit dysfunction, and hence may be compensatory to the primary pathology. Overall, the layer 4 network in the Fmr1-KO exhibits significant alterations in spike output in response to thalamocortical input and distorted sensory encoding. This developmental loss of layer 4 sensory encoding precision would contribute to subsequent developmental alterations in layer 4-to-layer 2/3 connectivity and plasticity observed in Fmr1-KO mice, and circuit dysfunction underlying sensory hypersensitivity.


Fmr1-KO brain
Notably, Wahlstrom-Helgren and Klyachko (2015) 1 have previously demonstrated aberrant FFI in the cortico-hippocampal feedforward circuit of P21 Fmr1-KO mice. The FFI alterations were synaptic pathway-specific and were attributed to altered presynaptic GABA B function. As in the present study, the authors reported a steady-state evoked EPSP broadening and an activity-dependent reduction in spike precision by the postsynaptic cell that required synaptic interactions in a two-pathway integration paradigm. We now demonstrate a role for altered postsynaptic intrinsic excitability in determining an aberrant functional response of the postsynaptic cell following presynaptic stimulation and extend the study to evaluate the ramification of altered FFI activity at the network level in somatosensory cortex. We also demonstrate distorted excitatory thalamocortical short-term plasticity contributes to the altered summation properties of layer 4 SCs. Short-term depressing synapses provide gain control and temporal filtering mechanisms that favour information transmission at lower frequencies 2,3 . We provide evidence for a shift in temporal and rate input discrimination in Fmr1-KO layer 4 that we attribute, in part, to stronger STP of depressing synapses. Notably, although these mechanisms could be linked by a common dependence on intracellular calcium signalling, Using modelling to identify causal vs. compensatory network-level contributions to network dysfunction Our modelling approaches have the potential to reduce cost, time and experimental animal usage. Supportive of this aim, as a companion to this paper we provide user-friendly simulation code with full documentation and usage tutorials to encourage non-specialists to experiment with the function of the circuits studied herein, as well as adapt them for their own use (see

Methods for hosting repository links).
To further address the question of causal vs. compensatory physiological disruptions, three main approaches will be necessary, firstly through physiological investigations spanning development time-points (this study and references 6,7 ) to evaluate the developmental evolution of features of physiological dysfunction, possibly in combination with chronic sensory manipulation (e.g. whisker deprivation) or carefully timed transient rescue interventions. This approach can be augmented by (secondly) analytical mathematical methods and (thirdly) numerical simulations. The derivation of exact analytical expressions governing network activity is notably challenging in the face of the complex, non-linear circuit dynamics 8,9 but reduced circuit or parametric statistical models (e.g. reference 10 ) offer powerful insight into which parameter(s) are dominant and which the network is robust to upon manipulation 10,11 .
Numerical simulation offers a complementary approach to studying network homeostasis and development 12,13 . These crucially provide predictions for the stochastic evolution of network structure and dynamics under different combinations of interventions, thereby comparing simulated network developmental trajectories from a common starting point. We introduce here a machine learning approach to classifying stimulus identity from multi-neuronal firing patterns. This could be extended to probe the evolving information capacity of model networks at different stages of in silico development. One promising approach is to compare a (model) circuit's observed information capacity to that of a theoretically optimal model, in which a tradeoff has been met for factors such as neural response correlation and response redundancy 14 .
The derivation of such an optimal circuit model could be hampered however by the animals' adoption of alternative behavioural strategies to extract relevant information from sensory input, as has been reported in Autistic individuals 15,16 .
Future experiments combining experimental manipulations of activity levels with model predictions will be able to elucidate which cellular parameters remain malleable and hence what pharmacological interventions may be more effective in restoring circuit function. These predictions can subsequently undergo rigorous testing by physiological experimentation. Such an approach would also be a valuable tool for examining the convergence of pathophysiology in disease mechanisms affecting circuit function across a range of genetic models.
Limitations and potential extensions of models used in this study Although our in silico models provide both mechanistic insight into and strong predictions about the quality of L4 circuit dysfunction in Fmr1-KO animals, it is important to recognise their limitations. Firstly, our two models lack 3D dendritic structure or detailed information about dendritic integration and thus only represent a first-order approximation of point summation at the soma. As such, potential changes to compartmentalised summation of thalamocortical input affecting active dendritic integration 17 or functional clustering of synaptic input could have been inadvertently overlooked. Moreover, we focused on genotype means of parameter sets grouped for simplicity by rational biophysical mechanisms. Consequently, we may have missed subtle antagonistic effects between collapsed parameters. For example, the changes in short-term plasticity we report in the experimental data are most likely attributable to presynaptic alterations that could potentially represent distinct changes in quantal release probability and/or failure rate that we capture in a single term in our phenomenological model of short-term depression. By providing the modelling code and analysis tools alongside this manuscript we encourage re-use of the simulations to both gain further intuitive insight into circuit function and genotype-dependent change and support novel analyses of similar datasets. One intriguing extension of our "rescue scenario" modelling approach could be to use dynamic clamp 18 to play in rescued Fmr1-KO currents into patch recordings from wildtype neurons (and vice versa). This could offer both validation of our findings and an opportunity to disentangle direct compensation vs. causative effects. For example, by both elevating the tone of Feed-forward inhibitory input to a wild-type SC and artificially reducing its leak conductance, one could examine whether an Fmr1-KO SC -like response emerged at a single-cell level, thereby removing changes to short-term plasticity from the potential causative mechanisms and isolating potential slower network-dependent compensations.
Modelling in this setting should not be used as a stand-alone tool. The above example represents one of several possible opportunities to integrate modelling as a predictive tool into a physiological experimental pipeline. Further detailed modelling efforts could include distributed inputs to the SC dendritic arbour and would ideally be accompanied by large-scale detailed Ca 2+ imaging of synaptic input to the SC neurons and systematic sampling of singlesynaptic input strength 19 .