Biology subjects

Gallitto, G.

Publications and source records attributed to Gallitto, G..

4 recordsLinked to original sources

bioRxiv2024

On the replicability of diffusion weighted MRI-based brain-behavior models

Establishing replicable inter-individual brain-wide associations is key to advancing our understanding of the crucial links between brain structure, function, and behavior, as well as applying this knowledge in clinical contexts. While the replicability and sample size requirements for anatomical and functional MRI-based brain-behavior associations have been extensively discussed recently, systematic replicability assessments are still lacking for diffusion-weighted imaging (DWI), despite it being the dominant non-invasive method to investigate white matter microstructure and structural connectivity. We report results of a comprehensive evaluation of the replicability of various DWI-based multivariate brain-behavior models. This evaluation is based on large-scale data from the Human Connectome Project, including five different DWI-based brain features (from fractional anisotropy to structural connectivity) and 58 different behavioral phenotypes. Our findings show an overall moderate replicability, with 24-31% of phenotypes replicable with sample sizes of fewer than 500. As DWI yields trait-like brain features, we restricted the analysis to trait-like phenotypes, such as cognitive and motor skills, and found much more promising replicability estimates, with 67-75% of these phenotypes replicable with n<500. Contrasting our empirical results to analytical replicability estimates substantiated that the replicability of DWI-based models is primarily a function of the true, unbiased effect size. Our work highlights the potential of DWI to produce replicable brain-behavior associations. However, it shows that achieving replicability with small-to-moderate samples requires stable, reliable and neurobiologically relevant target phenotypes. Our work highlights the potential of DWI to produce replicable brain-behavior associations, but only for stable, reliable and neurobiologically relevant target phenotypes. HIGHLIGHTSO_LIModerate replicability in DWI-based models: Overall replicability of DWI-based brain-behavior associations ranges from 24-31% with sample sizes under 500. C_LIO_LIImproved replicability for trait-like phenotypes: Trait-like phenotypes e.g., cognitive and motor skills exhibit higher replicability estimates of 67-75%, compared to state-like phenotypes such as emotion. C_LIO_LIEffect size as a key factor: Replicability is primarily influenced by the true, unbiased effect size, highlighting the importance of targeting stable and reliable phenotypes. C_LIO_LIPromise of -based multivariate associations: DWI-based brain-behaviour models should focus on phenotypes that display a sufficient temporal stability and test-retest reliability. C_LI

neuroscience↗

bioRxiv2024

Statistical learning of incidental perceptual regularities induces sensory conditioned cortical responses

Statistical learning of sensory patterns can lead to predictive neural processes enhancing stimulus perception and enabling fast deviancy detection. Predictive processes have been extensively demonstrated when environmental statistical regularities are relevant to task execution. Preliminary evidence indicates that statistical learning can even occur independently of task relevance and top-down attention, although the temporal profile and neural mechanisms underlying sensory predictions and error signals induced by statistical learning of incidental sensory regularities remain unclear. In our study, we adopted an implicit sensory conditioning paradigm that elicited the generation of specific perceptual priors in relation to task-irrelevant audio-visual associations, while recording Electroencephalography (EEG). Our results showed that learning of non-relevant but statistically interrelated neutral audio-visual stimuli resulted in early neural responses to predictive auditory stimuli conveying anticipatory signals of expected visual stimulus presence or absence, and in specific modulation of cortical responses to probabilistic visual stimulus presentation or omission. Pattern similarity analysis indicated that predictive auditory stimuli tended to resemble the response to expected visual stimulus presence or absence. Remarkably, Hierarchical Gaussian filter modeling estimating dynamic changes of prediction error signals in relation to differential probabilistic occurrences of audio-visual stimuli further demonstrated instantiation of predictive neural signals by showing distinct neural processing of prediction error in relation to violation of expected visual stimulus presence or absence. Overall, our findings indicated that statistical learning of non-salient and task-irrelevant perceptual regularities can induce the generation of neural priors at the time of predictive stimulus presentation, possibly conveying sensory-specific information of the predicted consecutive stimulus.

neuroscience↗

bioRxiv2023

External validation of machine learning models - registered models and adaptive sample splitting

Multivariate predictive models play a crucial role in enhancing our understanding of complex biological systems and in developing innovative, replicable tools for translational medical research. However, the complexity of machine learning methods and extensive data pre-processing and feature engineering pipelines can lead to overfitting and poor generalizability. An unbiased evaluation of predictive models necessitates external validation, which involves testing the finalized model on independent data. Despite its importance, external validation is often neglected in practice due to the associated costs. Here we propose that, for maximal credibility, model discovery and external validation should be separated by the public disclosure (e.g. pre-registration) of feature processing steps and model weights. Furthermore, we introduce a novel approach to optimize the trade-off between efforts spent on training and external validation in such studies. We show on data involving more than 3000 participants from four different datasets that, for any "sample size budget", the proposed adaptive splitting approach can successfully identify the optimal time to stop model discovery so that predictive performance is maximized without risking a low powered, and thus inconclusive, external validation. The proposed design and splitting approach (implemented in the Python package "AdaptiveSplit") may contribute to addressing issues of replicability, effect size inflation and generalizability in predictive modeling studies.

bioinformatics↗

bioRxiv2023

Connectome-Based Attractor Dynamics Guide Brain Activity in Rest, Task, and Disease

Functional brain connectivity has been instrumental in uncovering the large-scale organization of the brain and its relation to various behavioral and clinical phenotypes. Understanding how this functional architecture relates to the brains dynamic activity repertoire is an essential next step towards interpretable generative models of brain function. We propose functional connectivity-based Attractor Neural Networks (fcANNs), a theoretically inspired model of macro-scale brain dynamics, simulating recurrent activity flow among brain regions based on first principles of self-organization. In the fcANN framework, brain dynamics are understood in relation to attractor states; neurobiologically meaningful activity configurations that minimize the free energy of the system. We provide the first evidence that large-scale brain attractors - as reconstructed by fcANNs - exhibit an approximately orthogonal organization, which is a signature of the self-orthogonalization mechanism of the underlying theoretical framework of free-energy-minimizing attractor networks. Analyses of 7 distinct datasets demonstrate that fcANNs can accurately reconstruct and predict brain dynamics under a wide range of conditions, including resting and task states, and brain disorders. By establishing a formal link between connectivity and activity, fcANNs offer a simple and interpretable computational alternative to conventional descriptive analyses. Key PointsO_LIWe present a simple yet powerful generative computational model for large-scale brain dynamics C_LIO_LIBased on the theory of artificial attractor neural networks emerging from first principles of self-organization C_LIO_LIModel dynamics accurately reconstruct several characteristics of resting-state brain dynamics and confirm theoretical predictions of emergent attractor self-orthogonalization C_LIO_LIOur model captures both task-induced and pathological changes in brain activity C_LIO_LIfcANNs offer a simple and interpretable computational alternative to conventional descriptive analyses of brain function C_LI Project website (with interactive manuscript)https://pni-lab.github.io/connattractor

neuroscience↗