LMLaura Masaracchia

Laura Masaracchia

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Laura Masaracchia has a total of 11 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Hippocampal theta states are linked to reconfiguration of population activity with respect to behavioural context

    by Laura Masaracchia, Pablo Oyarzo, Felipe Fredes, Diego Vidaurre

    Neural activity reflects both external stimuli and the brain’s internal state, which shapes how information is processed and perceived. For example, in phase-precession in the hippocampus, the phase of theta oscillations affects the firing of single neurons (place cells) and its relation to the external world. Here, we examine a different form of relationship between oscillations and neural activity, where frequency and power of hippocampal oscillations, instead of phase, are associated with differential neural firing patterns at the population level. Using electrophysiological recordings of rats performing an odour-memory task and a data-driven model, we identify two distinct theta states: low-power-lower-theta and high-power-higher-theta. These states are associated to modulations in hippocampal neural ensemble activity representing trial outcome. Our findings suggest that power and frequency variations within theta oscillations may be linked to the reconfiguration of neural firing at the network level.

  • Open Access

    Long-Term Variability in Visual Processing versus Perceptual Stability

    by Laura Bock Paulsen, Laura Masaracchia, Francesca Fardo, Christine Ahrends, Diego Vidaurre

    Our brain is in constant change due to neural plasticity, but, still, our experience of the world feels relatively stable to us. Focusing on visual processing, we hypothesize that brain responses to stimuli may change over long periods of time, but in a way that is orthogonal to the dimensions that are relevant to stimulus category discrimination. To test this hypothesis, we acquired and analyzed a magnetoencephalography (MEG) dataset containing recordings from one female adult participant, with several scanning days spanning over 6 months. The participant passively attended to visual stimuli with the same stimulus presented within each session. We demonstrate that the specific scanning day can be predicted from the brain responses in a simple passive viewing paradigm, suggesting continuous temporal changes in neural activity over long time scales. However, information from one scanning day could be used to robustly decode the animacy of objects on a different scanning day, and importantly, decoding accuracy did not suffer with increasing time intervals between scans. That is, cross-decoding accuracy remained stable over months, despite between-scanning-day variability. These findings suggest that while processing of visual stimuli shows variability over long time scales, the core neural structure underlying object recognition remains stable and non-stimulus-specific. The results were validated in the open-access THINGS-MEG dataset, which employs a similar paradigm but covers a shorter longitudinal timespan. We find similar results across the four additional participants.

  • Open Access

    Frequency-specific theta states in the hippocampus are linked to reconfiguration of population activity with respect to behavioural context

    by Laura Masaracchia, Pablo Oyarzo, Felipe Fredes, Diego Vidaurre

    Neural activity reflects both external stimuli and the brain's internal state, which shapes how information is processed and perceived. An example of modulation of neural responses by network states is phase-precession in the hippocampus, where the phase of theta oscillations affects the firing of single neurons (place cells) and its relation to the external world. Here, we examine a different form of relationship between oscillations and neural activity, where frequency and power of the oscillation, instead of phase, are associated with differential neural firing patterns at the population level. We refer to this as ensemble pattern reconfiguration. To study this effect, we use electrophysiological recordings of rats performing an odour-memory (non-spatial) task. Using a data-driven model, we identified two distinct theta states: low-power-lower-theta (LPLT) and high-power-higher-theta (HPHT). Through regression and decoding analyses, we found that these states are associated to modulations in hippocampal neural ensemble activity representing trial outcome - though the effect did not consistently reach statistical significance in all our subjects. Our findings suggest that power and frequency variations within theta oscillations may be linked to the reconfiguration of neural firing at the network level, motivating further investigation in larger datasets.

  • Open Access

    Dissecting unsupervised learning through hidden Markov modeling in electrophysiological data

    by Laura Masaracchia, Felipe Fredes, Mark W. Woolrich, Diego Vidaurre

    Compared with classical supervised methods, unsupervised methods of analysis have the advantage to be freer of subjective biases. However, it is not always clear what aspects of the data these methods are most sensitive to, which complicates interpretation. Focusing on the hidden Markov model, commonly used to describe electrophysiological data, we explore in detail the nature of its estimates through simulations and real data examples, providing important insights about what to expect from these models.


  • Open Access

    Dissecting unsupervised learning through hidden Markov modelling in electrophysiological data

    by Laura Masaracchia, Felipe Fredes, Mark Woolrich, Diego Vidaurre

    Unsupervised, data-driven methods are commonly used in neuroscience to automatically decompose data into interpretable patterns. These patterns differ from one another depending on the assumptions of the models. How these assumptions affect specific data decompositions in practice, however, is often unclear, which hinders model applicability and interpretability. For instance, the hidden Markov model (HMM) automatically detects characteristic, recurring activity patterns (so-called states) from time series data. States are defined by a certain probability distribution, whose state-specific parameters are estimated from the data. But what specific features, from all of those that the data contain, do the states capture? That depends on the choice of probability distribution and on other model hyperparameters. Using both synthetic and real data, we aim at better characterising the behaviour of two HMM types that can be applied to electrophysiological data. Specifically, we study which differences in data features (such as frequency, amplitude or signal-to-noise ratio) are more salient to the models and therefore more likely to drive the state decomposition. Overall, we aim at providing guidance for an appropriate use of this type of analysis on one or two-channel neural electrophysiological data, and an informed interpretation of its results given the characteristics of the data and the purpose of the analysis.


  • Open Access

    Dissecting unsupervised learning through hidden Markov modelling in electrophysiological data

    by Laura Masaracchia, Felipe Fredes, Mark W. Woolrich, Diego Vidaurre

    Unsupervised, data-driven methods are commonly used in neuroscience to automatically decompose data into interpretable patterns. These patterns differ from one another depending on the assumptions of the models. How these assumptions affect specific data decompositions in practice, however, is often unclear, which hinders model applicability and interpretability. For instance, the hidden Markov model (HMM) automatically detects characteristic, recurring activity patterns (so-called states ) from time series data. States are defined by a certain probability distribution, whose state-specific parameters are estimated from the data. But what specific features, from all of those that the data contain, do the states capture? That depends on the choice of probability distribution and on other model hyperparameters. Using both synthetic and real data, we aim at better characterizing the behavior of two HMM types that can be applied to electrophysiological data. Specifically, we study which differences in data features (such as frequency, amplitude or signal-to-noise ratio) are more salient to the models and therefore more likely to drive the state decomposition. Overall, we aim at providing guidance for an appropriate use of this type of analysis on one or two-channel neural electrophysiological data, and an informed interpretation of its results given the characteristics of the data and the purpose of the analysis. NEW & NOTEWORTHY: Compared to classical supervised methods, unsupervised methods of analysis have the advantage to be freer of subjective biases. However, it is not always clear what aspects of the data these methods are most sensitive to, which complicates interpretation. Focusing on the Hidden Markov Model, commonly used to describe electrophysiological data, we explore in detail the nature of its estimates through simulations and real data examples, providing important insights about what to expect from these models.