Alex Roxin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Alex Roxin has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
- Klaus Wimmer
Centre de Recerca Matemàtica, Edifici C, Campus Bellaterra, Bellaterra, 080193, Barcelona, Spain
4 shared publications Federico Devalle
Centre de Recerca Matemàtica, Bellaterra, Spain
3 shared publications- Jaime de la Rocha
Institut d'Investigacions Biomèdiques August Pi i Sunyer, Barcelona, Spain
3 shared publications Albert Compte
Institut d'Investigacions Biomèdiques August Pi i Sunyer, Barcelona, Spain
2 shared publications- Genís Prat-Ortega
University of Pittsburgh
2 shared publications Gloria Cecchini
Centre de Recerca Matemàtica, Bellaterra, Spain
2 shared publications
Latest Publications
- Open Access
Random network structure stabilizes neural manifolds
Neuronal activity patterns change continuously over days and weeks, a phenomenon known as representational drift. Despite this, the geometric structure of population representations, namely the pairwise similarities between stimulus-evoked activity patterns, remains remarkably stable. How can ongoing changes in activity be consistent with stable representational similarity? We show that this is a generic consequence of random connectivity: in networks with random connectivity, output similarity is a monotonically increasing function of input similarity, independent of the specific connectivity pattern. Drift, whether driven by random synaptic turnover or Hebbian plasticity, merely transitions the network between random instantiations, leaving similarity intact. This extends to recurrent architectures and to deep neural networks, where continued training beyond performance saturation produces activity drift while preserving representational similarity. Although connectivity in the brain is not random, networks trained on high-dimensional inputs acquire connectivity that behaves statistically like a random projection, making these results broadly applicable to biological neural circuits.
Open AccessRandom network structure stabilizes neural manifolds
Neuronal activity patterns change continuously over days and weeks, a phenomenon known as representational drift. Despite this, the geometric structure of population representations, namely the pairwise similarities between stimulus-evoked activity patterns, remains remarkably stable. How can ongoing changes in activity be consistent with stable representational similarity? We show that this is a generic consequence of random connectivity: in networks with random connectivity, output similarity is a monotonically increasing function of input similarity, independent of the specific connectivity pattern. Drift, whether driven by random synaptic turnover or Hebbian plasticity, merely transitions the network between random instantiations, leaving similarity intact. This extends to recurrent architectures and to deep neural networks, where continued training beyond performance saturation produces activity drift while preserving representational similarity. Although connectivity in the brain is not random, networks trained on high-dimensional inputs acquire connectivity that behaves statistically like a random projection, making these results broadly applicable to biological neural circuits.
- Open Access
Locally balanced inhibition allows for robust learning of input-output associations in feedforward networks with Hebbian plasticity
In neural networks within the brain, the activity of a post-synaptic neuron is determined by the combined influence of many pre-synaptic neurons. This distributed processing enables mechanisms like Hebbian plasticity to associate sensory inputs with specific internal states, as seen in feedforward structures such as the CA1 region of the hippocampus. By modifying synaptic weights through Hebbian rules, sensory inputs can subsequently elicit outputs that consistently reflect their corresponding internal states. When input and output patterns are uncorrelated, this approach allows for the encoding of a large number of distinct associations, enabling efficient memory storage. Our study demonstrates a critical limitation when output patterns become weakly correlated with input patterns through the intrinsic feedforward network's connectivity. In these cases, the Hebbian rule preferentially strengthens synaptic weights shared across patterns, leading to a ``freezing'' of the network's structure. This results in highly correlated output patterns over time, effectively reducing the network's capacity to store diverse associations and limiting its flexibility in learning. To address this challenge, we find that including a mechanism of locally balanced inhibition, which has been shown to be a key feature of cortical circuits in-vivo, counteracts the undesired correlations between inputs and outputs. By dynamically regulating inhibitory input, locally balanced inhibition prevents the over-strengthening of shared weights, restoring the network's ability to maintain robust and flexible learning. This finding underscores the importance of inhibitory mechanisms in enabling efficient and adaptive information processing in neural circuits, offering insights into how biological networks maintain their remarkable capacity for associative learning.
Open AccessLocally balanced inhibition allows for robust learning of input-output associations in feedforward networks with Hebbian plasticity
In neural networks within the brain, the activity of a post-synaptic neuron is determined by the combined influence of many pre-synaptic neurons. This distributed processing enables mechanisms like Hebbian plasticity to associate sensory inputs with specific internal states, as seen in feedforward structures such as the CA1 region of the hippocampus. By modifying synaptic weights through Hebbian rules, sensory inputs can subsequently elicit outputs that consistently reflect their corresponding internal states. When input and output patterns are uncorrelated, this approach allows for the encoding of a large number of distinct associations, enabling efficient memory storage. Our study demonstrates a critical limitation when output patterns become weakly correlated with input patterns through the intrinsic feedforward network's connectivity. In these cases, the Hebbian rule preferentially strengthens synaptic weights shared across patterns, leading to a ``freezing'' of the network's structure. This results in highly correlated output patterns over time, effectively reducing the network's capacity to store diverse associations and limiting its flexibility in learning. To address this challenge, we find that including a mechanism of locally balanced inhibition, which has been shown to be a key feature of cortical circuits in-vivo, counteracts the undesired correlations between inputs and outputs. By dynamically regulating inhibitory input, locally balanced inhibition prevents the over-strengthening of shared weights, restoring the network's ability to maintain robust and flexible learning. This finding underscores the importance of inhibitory mechanisms in enabling efficient and adaptive information processing in neural circuits, offering insights into how biological networks maintain their remarkable capacity for associative learning.
- Open Access
Hebbian learning accounts for the effects of experience and novel exposure on representational drift in CA1 of hippocampus
Neuronal activity patterns slowly change over time even when sensory stimuli and animal behavior remain stable, a phenomenon known as representational drift. In area CA1 of the hippocampus, the amount of drift in the tuning of place cells on a familiar linear track is proportional to the time the animal spends exploring that track while the drift in cells' mean rate depends on the absolute time elapsed between sessions. Recently it was shown that exploration of a novel, enriched environment between sessions on the familiar track actually decreases the drift in tuning on that track, i.e. it stabilizes place fields compared to a baseline condition without the novel learning. This finding challenges computational models of drift in which new learning leads to overwriting and hence one would expect more drift, not less. Here we show that such models are indeed compatible with the observed findings as long as spatially-tuned input populations to CA1 which are active on the track versus the enriched environment are largely non-overlapping. Furthermore, in order to reproduce the findings, we must assume that the total amount of learning between the baseline and novel-exposure conditions is the same. Namely, the synaptic resources available for encoding patterns, or episodes, is fixed over a given period of time, but can be preferentially allocated to episodes of particular salience, such as the exploration of novel environments.
Open AccessHebbian learning accounts for the effects of experience and novel exposure on representational drift in CA1 of hippocampus
Neuronal activity patterns slowly change over time even when sensory stimuli and animal behavior remain stable, a phenomenon known as representational drift. In area CA1 of the hippocampus, the amount of drift in the tuning of place cells on a familiar linear track is proportional to the time the animal spends exploring that track while the drift in cells' mean rate depends on the absolute time elapsed between sessions. Recently it was shown that exploration of a novel, enriched environment between sessions on the familiar track actually decreases the drift in tuning on that track, i.e. it stabilizes place fields compared to a baseline condition without the novel learning. This finding challenges computational models of drift in which new learning leads to overwriting and hence one would expect more drift, not less. Here we show that such models are indeed compatible with the observed findings as long as spatially-tuned input populations to CA1 which are active on the track versus the enriched environment are largely non-overlapping. Furthermore, in order to reproduce the findings, we must assume that the total amount of learning between the baseline and novel-exposure conditions is the same. Namely, the synaptic resources available for encoding patterns, or episodes, is fixed over a given period of time, but can be preferentially allocated to episodes of particular salience, such as the exploration of novel environments.
- Open Access
- 30 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-20
Representational drift as the consequence of ongoing memory storage
Memory systems with biologically constrained synapses have been the topic of intense theoretical study for over thirty years. Perhaps the most fundamental and far-reaching finding from this work is that the storage of new memories implies the partial erasure of already-stored ones. This overwriting leads to a decorrelation of sensory-driven activity patterns over time, even if the input patterns remain similar. Representational drift (RD) should therefore be an expected and inevitable consequence of ongoing memory storage. We tested this hypothesis by fitting a network model to data from long-term chronic calcium imaging experiments in mouse hippocampus. Synaptic turnover in the model inputs, consistent with the ongoing encoding of new activity patterns, accounted for the observed statistics of RD. This mechanism also provides a parsimonious explanation for the diverse effects of experience on drift found in experiment. Our results suggest that RD should be observed wherever neuronal circuits are involved in a process of ongoing learning or memory storage.
- 30 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-20
Open AccessRepresentational drift as the consequence of ongoing memory storage
Memory systems with biologically constrained synapses have been the topic of intense theoretical study for over thirty years. Perhaps the most fundamental and far-reaching finding from this work is that the storage of new memories implies the partial erasure of already-stored ones. This overwriting leads to a decorrelation of sensory-driven activity patterns over time, even if the input patterns remain similar. Representational drift (RD) should therefore be an expected and inevitable consequence of ongoing memory storage. We tested this hypothesis by fitting a network model to data from long-term chronic calcium imaging experiments in mouse hippocampus. Synaptic turnover in the model inputs, consistent with the ongoing encoding of new activity patterns, accounted for the observed statistics of RD. This mechanism also provides a parsimonious explanation for the diverse effects of experience on drift found in experiment. Our results suggest that RD should be observed wherever neuronal circuits are involved in a process of ongoing learning or memory storage.
- 11 December 2024
- Published by Springer Nature in Journal of Computational Neuroscience
- Vol. 53 (1) , 9-23
How plasticity shapes the formation of neuronal assemblies driven by oscillatory and stochastic inputs
Synaptic connections in neuronal circuits are modulated by pre- and post-synaptic spiking activity. Previous theoretical work has studied how such Hebbian plasticity rules shape network connectivity when firing rates are constant, or slowly varying in time. However, oscillations and fluctuations, which can arise through sensory inputs or intrinsic brain mechanisms, are ubiquitous in neuronal circuits. Here we study how oscillatory and fluctuating inputs shape recurrent network connectivity given a temporally asymmetric plasticity rule. We do this analytically using a separation of time scales approach for pairs of neurons, and then show that the analysis can be extended to understand the structure in large networks. In the case of oscillatory inputs, the resulting network structure is strongly affected by the phase relationship between drive to different neurons. In large networks, distributed phases tend to lead to hierarchical clustering. The analysis for stochastic inputs reveals a rich phase plane in which there is multistability between different possible connectivity motifs. Our results may be of relevance for understanding the effect of sensory-driven inputs, which are by nature time-varying, on synaptic plasticity, and hence on learning and memory.
- 11 December 2024
- Published by Springer Nature in Journal of Computational Neuroscience
- Vol. 53 (1) , 9-23
How plasticity shapes the formation of neuronal assemblies driven by oscillatory and stochastic inputs
Synaptic connections in neuronal circuits are modulated by pre- and post-synaptic spiking activity. Previous theoretical work has studied how such Hebbian plasticity rules shape network connectivity when firing rates are constant, or slowly varying in time. However, oscillations and fluctuations, which can arise through sensory inputs or intrinsic brain mechanisms, are ubiquitous in neuronal circuits. Here we study how oscillatory and fluctuating inputs shape recurrent network connectivity given a temporally asymmetric plasticity rule. We do this analytically using a separation of time scales approach for pairs of neurons, and then show that the analysis can be extended to understand the structure in large networks. In the case of oscillatory inputs, the resulting network structure is strongly affected by the phase relationship between drive to different neurons. In large networks, distributed phases tend to lead to hierarchical clustering. The analysis for stochastic inputs reveals a rich phase plane in which there is multistability between different possible connectivity motifs. Our results may be of relevance for understanding the effect of sensory-driven inputs, which are by nature time-varying, on synaptic plasticity, and hence on learning and memory.
- Open Access
Representational drift as the consequence of ongoing memory storage
Memory systems with biologically constrained synapses have been the topic of intense theoretical study for over thirty years. Perhaps the most fundamental and far-reaching finding from this work is that the storage of new memories implies the partial erasure of already-stored ones. This overwriting leads to a decorrelation of sensory-driven activity patterns over time, even if the input patterns remain similar. Representational drift (RD) should therefore be an expected and inevitable consequence of ongoing memory storage. We tested this hypothesis by fitting a network model to data from long-term chronic calcium imaging experiments in mouse hippocampus. Synaptic turnover in the model inputs, consistent with the ongoing encoding of new activity patterns, accounted for the observed statistics of RD. This mechanism also provides a parsimonious explanation for the diverse effects of experience on drift found in experiment. Our results suggest that RD should be observed wherever neuronal circuits are involved in a process of ongoing learning or memory storage.
Open AccessRepresentational drift as the consequence of ongoing memory storage
Memory systems with biologically constrained synapses have been the topic of intense theoretical study for over thirty years. Perhaps the most fundamental and far-reaching finding from this work is that the storage of new memories implies the partial erasure of already-stored ones. This overwriting leads to a decorrelation of sensory-driven activity patterns over time, even if the input patterns remain similar. Representational drift (RD) should therefore be an expected and inevitable consequence of ongoing memory storage. We tested this hypothesis by fitting a network model to data from long-term chronic calcium imaging experiments in mouse hippocampus. Synaptic turnover in the model inputs, consistent with the ongoing encoding of new activity patterns, accounted for the observed statistics of RD. This mechanism also provides a parsimonious explanation for the diverse effects of experience on drift found in experiment. Our results suggest that RD should be observed wherever neuronal circuits are involved in a process of ongoing learning or memory storage.
- Open Access
- 5 November 2023
- Published by Elsevier in IBRO Neuroscience Reports
- Vol. 15, S777
NEURAL NETWORK MECHANISMS UNDERLYING POST-DECISION BIASES
- 5 November 2023
- Published by Elsevier in IBRO Neuroscience Reports
- Vol. 15, S777
Open AccessNEURAL NETWORK MECHANISMS UNDERLYING POST-DECISION BIASES
- Open Access
- 5 November 2023
- Published by Elsevier in IBRO Neuroscience Reports
- Vol. 15, S806
RAPID MEMORY ENCODING IN A RECURRENT NETWORK MODEL WITH BEHAVIORAL TIME SCALE SYNAPTIC PLASTICITY
- 5 November 2023
- Published by Elsevier in IBRO Neuroscience Reports
- Vol. 15, S806
Open AccessRAPID MEMORY ENCODING IN A RECURRENT NETWORK MODEL WITH BEHAVIORAL TIME SCALE SYNAPTIC PLASTICITY
- Open Access
- 25 August 2023
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 19 (8) , e1011139
Rapid memory encoding in a recurrent network model with behavioral time scale synaptic plasticity
Episodic memories are formed after a single exposure to novel stimuli. The plasticity mechanisms underlying such fast learning still remain largely unknown. Recently, it was shown that cells in area CA1 of the hippocampus of mice could form or shift their place fields after a single traversal of a virtual linear track. In-vivo intracellular recordings in CA1 cells revealed that previously silent inputs from CA3 could be switched on when they occurred within a few seconds of a dendritic plateau potential (PP) in the post-synaptic cell, a phenomenon dubbed Behavioral Time-scale Plasticity (BTSP). A recently developed computational framework for BTSP in which the dynamics of synaptic traces related to the pre-synaptic activity and post-synaptic PP are explicitly modelled, can account for experimental findings. Here we show that this model of plasticity can be further simplified to a 1D map which describes changes to the synaptic weights after a single trial. We use a temporally symmetric version of this map to study the storage of a large number of spatial memories in a recurrent network, such as CA3. Specifically, the simplicity of the map allows us to calculate the correlation of the synaptic weight matrix with any given past environment analytically. We show that the calculated memory trace can be used to predict the emergence and stability of bump attractors in a high dimensional neural network model endowed with BTSP.
- 25 August 2023
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 19 (8) , e1011139
Open AccessRapid memory encoding in a recurrent network model with behavioral time scale synaptic plasticity
Episodic memories are formed after a single exposure to novel stimuli. The plasticity mechanisms underlying such fast learning still remain largely unknown. Recently, it was shown that cells in area CA1 of the hippocampus of mice could form or shift their place fields after a single traversal of a virtual linear track. In-vivo intracellular recordings in CA1 cells revealed that previously silent inputs from CA3 could be switched on when they occurred within a few seconds of a dendritic plateau potential (PP) in the post-synaptic cell, a phenomenon dubbed Behavioral Time-scale Plasticity (BTSP). A recently developed computational framework for BTSP in which the dynamics of synaptic traces related to the pre-synaptic activity and post-synaptic PP are explicitly modelled, can account for experimental findings. Here we show that this model of plasticity can be further simplified to a 1D map which describes changes to the synaptic weights after a single trial. We use a temporally symmetric version of this map to study the storage of a large number of spatial memories in a recurrent network, such as CA3. Specifically, the simplicity of the map allows us to calculate the correlation of the synaptic weight matrix with any given past environment analytically. We show that the calculated memory trace can be used to predict the emergence and stability of bump attractors in a high dimensional neural network model endowed with BTSP.
- Open Access
Rapid memory encoding in a recurrent network model with behavioral time scale synaptic plasticity
Episodic memories are formed after a single exposure to novel stimuli. The plasticity mechanisms underlying such fast learning still remain largely unknown. Recently, it was shown that cells in area CA1 of the hippocampus of mice could form or shift their place fields after a single traversal of a virtual linear track. In-vivo intracellular recordings in CA1 cells revealed that previously silent inputs from CA3 could be switched on when they occurred within a few seconds of a dendritic plateau potential (PP) in the post-synaptic cell, a phenomenon dubbed Behavioral Time-scale Plasticity (BTSP). A recently developed computational framework for BTSP in which the dynamics of synaptic traces related to the pre-synaptic activity and post-synaptic PP are explicitly modelled, can account for experimental findings. Here we show that this model of plasticity can be further simplified to simple map which describes changes to the synaptic weights after a single trial. The map can be solved analytically and provides a quantitative fit to experimental data from CA1. We use a temporally symmetric version of this map to study the storage of a large number of spatial memories in a recurrent network, such as CA3. Specifically, the simplicity of the map allows us to calculate the correlation of the synaptic weight matrix with any given past environment analytically. We show that the calculated memory trace can be used to predict the emergence and stability of bump attractors in a high dimensional neural network model endowed with BTSP. Author summary A recently discovered form of in-vivo plasticity, called Behavioral Time-scale Plasticity (BTSP), leads to the generation of a place cell in CA1 from a previously silent cell after a single intracellular plateau potential (PP). We show that this one-shot learning process is well-described by a 1D map, which updates the synaptic weight matrix after the PP. We use the map to study the storage of spatial memories in a recurrent network, such as CA3. The map allows us to calculate the correlation of the weight matrix with all past explored environments analytically. When only a small fraction of cells are place cells in any given environment, we show that the full dynamics of a recurrent network endowed with BTSP is equivalent to its projection onto a series of weakly interacting manifolds, one for each environment. Interestingly, this weak interaction, which takes the form of quenched variability, actually enhances the memory capacity by stabilizing remote memories which otherwise would be unretrievable.
Open AccessRapid memory encoding in a recurrent network model with behavioral time scale synaptic plasticity
Episodic memories are formed after a single exposure to novel stimuli. The plasticity mechanisms underlying such fast learning still remain largely unknown. Recently, it was shown that cells in area CA1 of the hippocampus of mice could form or shift their place fields after a single traversal of a virtual linear track. In-vivo intracellular recordings in CA1 cells revealed that previously silent inputs from CA3 could be switched on when they occurred within a few seconds of a dendritic plateau potential (PP) in the post-synaptic cell, a phenomenon dubbed Behavioral Time-scale Plasticity (BTSP). A recently developed computational framework for BTSP in which the dynamics of synaptic traces related to the pre-synaptic activity and post-synaptic PP are explicitly modelled, can account for experimental findings. Here we show that this model of plasticity can be further simplified to simple map which describes changes to the synaptic weights after a single trial. The map can be solved analytically and provides a quantitative fit to experimental data from CA1. We use a temporally symmetric version of this map to study the storage of a large number of spatial memories in a recurrent network, such as CA3. Specifically, the simplicity of the map allows us to calculate the correlation of the synaptic weight matrix with any given past environment analytically. We show that the calculated memory trace can be used to predict the emergence and stability of bump attractors in a high dimensional neural network model endowed with BTSP. Author summary A recently discovered form of in-vivo plasticity, called Behavioral Time-scale Plasticity (BTSP), leads to the generation of a place cell in CA1 from a previously silent cell after a single intracellular plateau potential (PP). We show that this one-shot learning process is well-described by a 1D map, which updates the synaptic weight matrix after the PP. We use the map to study the storage of spatial memories in a recurrent network, such as CA3. The map allows us to calculate the correlation of the weight matrix with all past explored environments analytically. When only a small fraction of cells are place cells in any given environment, we show that the full dynamics of a recurrent network endowed with BTSP is equivalent to its projection onto a series of weakly interacting manifolds, one for each environment. Interestingly, this weak interaction, which takes the form of quenched variability, actually enhances the memory capacity by stabilizing remote memories which otherwise would be unretrievable.