Guillermo Cecchi

Alzheimers Disease Brain Phenotypes are Age-dependent

This paper shows that age information is essential for Alzheimer's Disease detection using brain MRI, outperforming age-invariant methods and revealing flaws in brain-age gap as a biomarker.

Multimodal fusion of brain signals for robust prediction of psychosis transition

The paper uses multimodal machine learning to predict psychosis conversion and finds that integrating multiple data types improves prediction accuracy.

Speak and you shall predict: speech at initial cocaine abstinence as a biomarker of long-term drug use behavior

This paper shows that speech analysis using natural language processing can predict long-term cocaine use outcomes better than traditional measures in initially abstinent individuals with cocaine use disorder.

Expansive Linguistic Representations to Predict Interpretable Odor Mixture Discriminability

The paper uses semantic descriptors from language to predict odor mixture discriminability, outperforming existing methods by leveraging human perception data.

Predicting human decision making in psychological tasks with recurrent neural networks

This paper uses LSTM networks to predict human decision-making trajectories in psychological tasks, showing better performance than existing methods in both single- and multi-agent scenarios.

MVD-Fuse: Detection of White Matter Degeneration via Multi-View Learning of Diffusion Microstructure

MVD-Fuse integrates diffusion MRI models to detect white matter degeneration, showing improved classification of pre-HD and healthy controls with MKL fusion.

Speech-based identification of L-DOPA ON/OFF state in Parkinson’s Disease subjects

This study uses speech analysis to classify Parkinson's disease patients' L-DOPA medication states with high accuracy.

Detection of Amyotrophic Lateral Sclerosis (ALS) via Acoustic Analysis

This paper uses acoustic speech analysis to detect and assess ALS progression, achieving 79% accuracy for males and 83% for females using a mobile app dataset.

Brain and psychological determinants of placebo pill response in chronic pain patients

This study shows that placebo pill response in chronic pain patients depends on brain structure, functional connectivity, and psychological factors, with machine learning predicting response magnitude.

Predicting natural language descriptions of smells

The paper predicts perceptual ratings for odorants using natural language semantics and chemoinformatic features, achieving higher accuracy for more descriptors than previous methods.

Learning Nonlinear Brain Dynamics: van der Pol Meets LSTM

The paper introduces a nonlinear differential equation model for brain dynamics using van der Pol oscillators and improves LSTM performance through model-specific data augmentation.

Learning to read the imprints of consciousness on global brain dynamics: an application to intra-operative monitoring of anesthesia

This paper develops a method to monitor anesthesia depth by analyzing brain dynamics, showing it can distinguish between awake and unconscious states with high accuracy.

Reverse-engineering human olfactory perception from chemical features of odor molecules

The paper demonstrates that machine learning can accurately predict odor perception from molecular features, enabling reverse-engineering of smell.