Lorenzo Gaetano Amato
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Lorenzo Gaetano Amato has a total of 52 co-authors. Here are the top co-authors by number of shared publications:
Alberto Arturo Vergani
The Biorobotics Institute, Sant’Anna School of Advanced Studies, Pisa, Italy
5 shared publications- Alberto Mazzoni
The Biorobotics Institute, Sant’Anna School of Advanced Studies, Pisa, Italy
5 shared publications Michael Lassi
The BioRobotics Institute and Department of Excellence in Robotics and AI, Scuola Superiore Sant’Anna, Pisa, Italy
5 shared publicationsAntonello Grippo
Neurophysiology Unit, Careggi University Hospital, Florence, Italy
4 shared publicationsBenedetta Nacmias
Department of Neuroscience, Psychology, Drug Research and Child Health, University of Florence, Florence, Italy
4 shared publicationsCarlo Fabbiani
IRCCS Fondazione Don Carlo Gnocchi, Florence, Italy
4 shared publications
Latest Publications
- Open Access
Normative EEG Effective Connectivity as a neural marker of Autism Spectrum Disorder
Autism spectrum disorder (ASD) is one of the most common neurodevelopmental disorders. The absence of a clear aetiology strengthens the need to identify quantitative biomarkers capable of diagnosing the condition and assessing the severity of symptoms. We implemented a normative modelling framework based on EEG effective connectivity (EC), leveraging recordings collected in three centers, in typical development (TD) and ASD participants with age spanning from 5 to 19 years. Deviations from the normative evolution were tested as ASD biomarkers, compared with standard EC metrics. Normative EC metrics supported accurate classification of ASD vs TD participants (0.82 AUC), reaching 0.95 AUC when combined with behavioral subscales, significantly outperforming a classification based on behavioral data alone. In ASD participants, selected metrics in the Temporo-Parietal-Frontal network predicted social responsiveness scale (SRS) values with high significance (SRS Total R 2 =0.42, p=0.0005). These results suggest that normative EC values constitute robust and generalizable ASD biomarkers.
Open AccessNormative EEG Effective Connectivity as a neural marker of Autism Spectrum Disorder
Autism spectrum disorder (ASD) is one of the most common neurodevelopmental disorders. The absence of a clear aetiology strengthens the need to identify quantitative biomarkers capable of diagnosing the condition and assessing the severity of symptoms. We implemented a normative modelling framework based on EEG effective connectivity (EC), leveraging recordings collected in three centers, in typical development (TD) and ASD participants with age spanning from 5 to 19 years. Deviations from the normative evolution were tested as ASD biomarkers, compared with standard EC metrics. Normative EC metrics supported accurate classification of ASD vs TD participants (0.82 AUC), reaching 0.95 AUC when combined with behavioral subscales, significantly outperforming a classification based on behavioral data alone. In ASD participants, selected metrics in the Temporo-Parietal-Frontal network predicted social responsiveness scale (SRS) values with high significance (SRS Total R 2 =0.42, p=0.0005). These results suggest that normative EC values constitute robust and generalizable ASD biomarkers.
- Open Access
Brain Network Excitability Predicts Clinical Severity in Multiple Sclerosis
Comprehensive biomarkers of multiple sclerosis (MS) capable of simultaneously diagnosing the condition, capturing symptom severity and predicting treatment efficacy remain elusive. Although several studies have highlighted the pivotal role played by demyelinating lesions in determining MS structural pathology, their relationship with symptom severity is limited. Here, we combined personalized computational brain modeling with magnetoencephalography (MEG) recordings from 17 MS patients and 20 healthy controls (CTR) to derive personalized brain network excitability parameters, which we tested as MS biomarkers. Personalized parameters discriminated between CTR and MS participants with high accuracy, also classifying between progressing and remitting MS patients. Notably, they also predicted MS clinical scales across multiple domains. In all clinical tasks, personalized parameters consistently outperformed standard clinical measures and total lesion loads. Together, these results highlight the potential of personalized brain modelling in deriving integrative MS biomarkers, capable of simultaneously identifying the condition, classifying MS subtypes and predicting symptom severity.
Open AccessBrain Network Excitability Predicts Clinical Severity in Multiple Sclerosis
Comprehensive biomarkers of multiple sclerosis (MS) capable of simultaneously diagnosing the condition, capturing symptom severity and predicting treatment efficacy remain elusive. Although several studies have highlighted the pivotal role played by demyelinating lesions in determining MS structural pathology, their relationship with symptom severity is limited. Here, we combined personalized computational brain modeling with magnetoencephalography (MEG) recordings from 17 MS patients and 20 healthy controls (CTR) to derive personalized brain network excitability parameters, which we tested as MS biomarkers. Personalized parameters discriminated between CTR and MS participants with high accuracy, also classifying between progressing and remitting MS patients. Notably, they also predicted MS clinical scales across multiple domains. In all clinical tasks, personalized parameters consistently outperformed standard clinical measures and total lesion loads. Together, these results highlight the potential of personalized brain modelling in deriving integrative MS biomarkers, capable of simultaneously identifying the condition, classifying MS subtypes and predicting symptom severity.
- Open Access
Can we trust subthalamic local field potentials? Geometric and dynamical constraints on the interpretation of extracellular recordings
Local field potentials (LFPs) are widely interpreted as readouts of population synaptic activity, an assumption derived almost entirely from cortical recordings. Whether these principles extend to subcortical structures remains unclear. We address this question in the subthalamic nucleus (STN), where LFPs are routinely recorded and used to guide adaptive deep brain stimulation for Parkinson’s disease, using a biophysically detailed population model benchmarked against patient microelectrode recordings. As in the cortex, STN extracellular potentials were dominated by synaptic currents. Differently from the cortex, however, LFPs could not be reliably predicted from these currents or other average population quantities. This dissociation arises from the STN symmetric neuronal morphology and lack of recurrent connectivity, which promote destructive interference among single-neuron contributions, decoupling the LFP from population-level dynamics. This decoupling was not absolute: pathological beta synchrony restored a robust synapse-LFP relationship by consistent underlying dynamics, while the aperiodic slope of the power spectral density tracked STN neuronal morphology, firing rate, and excitatory-inhibitory balance. Together, these findings challenge the prevailing view of LFPs as universal readouts of population activity. Our results show that the interpretability of extracellular signals depends critically on neuronal morphology and synchronization state, and provide a mechanistic framework for the use of STN LFPs as biomarkers in adaptive deep brain stimulation for Parkinson’s disease.
Open AccessCan we trust subthalamic local field potentials? Geometric and dynamical constraints on the interpretation of extracellular recordings
Local field potentials (LFPs) are widely interpreted as readouts of population synaptic activity, an assumption derived almost entirely from cortical recordings. Whether these principles extend to subcortical structures remains unclear. We address this question in the subthalamic nucleus (STN), where LFPs are routinely recorded and used to guide adaptive deep brain stimulation for Parkinson’s disease, using a biophysically detailed population model benchmarked against patient microelectrode recordings. As in the cortex, STN extracellular potentials were dominated by synaptic currents. Differently from the cortex, however, LFPs could not be reliably predicted from these currents or other average population quantities. This dissociation arises from the STN symmetric neuronal morphology and lack of recurrent connectivity, which promote destructive interference among single-neuron contributions, decoupling the LFP from population-level dynamics. This decoupling was not absolute: pathological beta synchrony restored a robust synapse-LFP relationship by consistent underlying dynamics, while the aperiodic slope of the power spectral density tracked STN neuronal morphology, firing rate, and excitatory-inhibitory balance. Together, these findings challenge the prevailing view of LFPs as universal readouts of population activity. Our results show that the interpretability of extracellular signals depends critically on neuronal morphology and synchronization state, and provide a mechanistic framework for the use of STN LFPs as biomarkers in adaptive deep brain stimulation for Parkinson’s disease.
- Open Access
- 11 June 2026
- Published by Public Library of Science (PLoS) in PLOS Digital Health
- Vol. 5 (6) , e0001445
A digital twin approach for simultaneous reconstruction of brain anatomy and dynamics from neural data
Brain structure plays a pivotal role in shaping neural dynamics. Current models lack the anatomical and functional resolution needed to integrate whole-brain structure and dynamics within a unified computational framework. Here, we introduce the FEDE (high FidElity Digital brain modEl) pipeline, generating anatomically accurate brain digital twins from imaging data. Combining advanced techniques of finite-element analysis and biophysical modeling, FEDE reconstructs multi-scale brain structure with high spatial resolution, while also replicating whole-brain neural activity. We demonstrated FEDE’s application by creating the first brain digital twin of a toddler with autism spectrum disorder (ASD). Through parameter optimization, FEDE replicated experimental neural activity while reconstructing multi-scale structural features ranging from whole-brain connectivity to synaptic timescales. FEDE estimated possible patient-specific anomalies in synaptic transmission, consistent with ASD pathophysiology. Our pipeline represents a significant leap forward in brain modeling, paving the way for effective applications of digital twins in experimental and clinical settings.
- 11 June 2026
- Published by Public Library of Science (PLoS) in PLOS Digital Health
- Vol. 5 (6) , e0001445
Open AccessA digital twin approach for simultaneous reconstruction of brain anatomy and dynamics from neural data
Brain structure plays a pivotal role in shaping neural dynamics. Current models lack the anatomical and functional resolution needed to integrate whole-brain structure and dynamics within a unified computational framework. Here, we introduce the FEDE (high FidElity Digital brain modEl) pipeline, generating anatomically accurate brain digital twins from imaging data. Combining advanced techniques of finite-element analysis and biophysical modeling, FEDE reconstructs multi-scale brain structure with high spatial resolution, while also replicating whole-brain neural activity. We demonstrated FEDE’s application by creating the first brain digital twin of a toddler with autism spectrum disorder (ASD). Through parameter optimization, FEDE replicated experimental neural activity while reconstructing multi-scale structural features ranging from whole-brain connectivity to synaptic timescales. FEDE estimated possible patient-specific anomalies in synaptic transmission, consistent with ASD pathophysiology. Our pipeline represents a significant leap forward in brain modeling, paving the way for effective applications of digital twins in experimental and clinical settings.
- 28 January 2026
- Published by SAGE Publications in Journal of Alzheimer’s Disease
- Vol. 110 (2) , 728-745
Towards the development of a management protocol for subjective cognitive decline: Insights from a cross-sectional and longitudinal analysis of multimodal data from a memory clinic
Background: Subjective cognitive decline (SCD) represents the first early symptomatic stage of Alzheimer's disease (AD). Objective: We aimed to investigate the relationships between features in SCD and to assess the importance of these features in the future development of dementia to inform a targeted management protocol. Methods: 440 SCD patients underwent neurological and neuropsychological assessments, MRI scans, APOE genotyping, and AD biomarker evaluations. Patients were followed for a median of 10 years. Relationships among features were first assessed univariately, focusing on differences across stratified subgroups. To capture multivariate associations, we applied network analysis using a Markov Random Field. Finally, baseline features were related to dementia progression using an XGboost machine learning model. Results: Women comprising 68.9% of the cohort, were generally younger at onset, had lower APOE ε4 prevalence, and differed in neuropsychological performance compared to men. Older patients (age >60) exhibited a higher prevalence of APOE ε4 and cerebral small vessel disease. Patients with depressive symptoms demonstrated lower cognitive performance across multiple domains. Network analysis indicated complex interconnections among gender, cognitive reserve, SCD severity, and depressive symptoms. The XGboost model achieved 74% accuracy in predicting progression to dementia, identifying age at onset, mini-mental state examination scores, and APOE genotype as the most predictive factors. Conclusions: This study highlights the role of age, gender, APOE genotype, and depressive symptoms in the presentation and progression of cognitive decline. By identifying key predictive features, we propose a personalized management protocol aimed at optimizing care for individuals with SCD. Trial registration number: NCT05569083, registration date: 2019-05-30
- 28 January 2026
- Published by SAGE Publications in Journal of Alzheimer’s Disease
- Vol. 110 (2) , 728-745
Towards the development of a management protocol for subjective cognitive decline: Insights from a cross-sectional and longitudinal analysis of multimodal data from a memory clinic
Background: Subjective cognitive decline (SCD) represents the first early symptomatic stage of Alzheimer's disease (AD). Objective: We aimed to investigate the relationships between features in SCD and to assess the importance of these features in the future development of dementia to inform a targeted management protocol. Methods: 440 SCD patients underwent neurological and neuropsychological assessments, MRI scans, APOE genotyping, and AD biomarker evaluations. Patients were followed for a median of 10 years. Relationships among features were first assessed univariately, focusing on differences across stratified subgroups. To capture multivariate associations, we applied network analysis using a Markov Random Field. Finally, baseline features were related to dementia progression using an XGboost machine learning model. Results: Women comprising 68.9% of the cohort, were generally younger at onset, had lower APOE ε4 prevalence, and differed in neuropsychological performance compared to men. Older patients (age >60) exhibited a higher prevalence of APOE ε4 and cerebral small vessel disease. Patients with depressive symptoms demonstrated lower cognitive performance across multiple domains. Network analysis indicated complex interconnections among gender, cognitive reserve, SCD severity, and depressive symptoms. The XGboost model achieved 74% accuracy in predicting progression to dementia, identifying age at onset, mini-mental state examination scores, and APOE genotype as the most predictive factors. Conclusions: This study highlights the role of age, gender, APOE genotype, and depressive symptoms in the presentation and progression of cognitive decline. By identifying key predictive features, we propose a personalized management protocol aimed at optimizing care for individuals with SCD. Trial registration number: NCT05569083, registration date: 2019-05-30
- Open Access
- 15 January 2026
- Published by Springer Nature in Communications Medicine
- Vol. 6 (1) , 30
Digital twins support cross-modal and cross-centric classification of mild cognitive impairment
Background Neural recordings capture crucial pathophysiological processes along the dementia continuum. However, cross-center variability in recording techniques and paradigms limit their generalizability and diagnostic power, preventing clinical use. We here propose a computational approach enabling cross-center classification even in the presence of completely different clinical pipelines. Methods We leveraged a digital twin model to derive digital biomarkers linking neurodegeneration mechanisms to alterations in neural activity across multiple recording modalities. We tested the generalizability of digital biomarkers through cross-center classification of Mild Cognitive Impairment (MCI) and healthy subjects in two independent clinics. The two datasets presented different recording techniques (EEG and MEG), preprocessing modalities, recruitment criteria and diagnostic guidelines. Digital biomarkers derived from one clinic were tested for classifying patients in the other clinic and vice versa employing a transfer learning approach. Results Digital biomarkers outperform standard biomarkers in the MCI vs healthy classification in both separate datasets (83% vs 58% for EEG dataset and 75% vs 68% for MEG dataset). Moreover, they achieve accurate and consistent cross-center classification (77–78% accuracy), while standard biomarkers perform poorly in the generalization attempt (56–65%). Additionally, digital biomarkers reliably predict global cognitive status across clinics across both datasets ( p < 0.01), while standard biomarkers present no correlation. Conclusions Digital biomarkers generalize across recording techniques and datasets, enabling a cross-modal and cross-center classification of a patient’s condition. These biomarkers offer a robust measure of patient-specific neurodegeneration, mapping neural recordings anomalies into a common framework of underlying structural alterations. The vast differences between the two datasets support the applicability of this approach also in the presence of high inter-center variability.
- 15 January 2026
- Published by Springer Nature in Communications Medicine
- Vol. 6 (1) , 30
Open AccessDigital twins support cross-modal and cross-centric classification of mild cognitive impairment
Background Neural recordings capture crucial pathophysiological processes along the dementia continuum. However, cross-center variability in recording techniques and paradigms limit their generalizability and diagnostic power, preventing clinical use. We here propose a computational approach enabling cross-center classification even in the presence of completely different clinical pipelines. Methods We leveraged a digital twin model to derive digital biomarkers linking neurodegeneration mechanisms to alterations in neural activity across multiple recording modalities. We tested the generalizability of digital biomarkers through cross-center classification of Mild Cognitive Impairment (MCI) and healthy subjects in two independent clinics. The two datasets presented different recording techniques (EEG and MEG), preprocessing modalities, recruitment criteria and diagnostic guidelines. Digital biomarkers derived from one clinic were tested for classifying patients in the other clinic and vice versa employing a transfer learning approach. Results Digital biomarkers outperform standard biomarkers in the MCI vs healthy classification in both separate datasets (83% vs 58% for EEG dataset and 75% vs 68% for MEG dataset). Moreover, they achieve accurate and consistent cross-center classification (77–78% accuracy), while standard biomarkers perform poorly in the generalization attempt (56–65%). Additionally, digital biomarkers reliably predict global cognitive status across clinics across both datasets ( p < 0.01), while standard biomarkers present no correlation. Conclusions Digital biomarkers generalize across recording techniques and datasets, enabling a cross-modal and cross-center classification of a patient’s condition. These biomarkers offer a robust measure of patient-specific neurodegeneration, mapping neural recordings anomalies into a common framework of underlying structural alterations. The vast differences between the two datasets support the applicability of this approach also in the presence of high inter-center variability.
- Open Access
- 31 May 2025
- Published by Springer Nature in Alzheimer's Research & Therapy
- Vol. 17 (1) , 1-16
Digital twins and non-invasive recordings enable early diagnosis of Alzheimer’s disease
The diagnosis of Alzheimer’s disease (AD) in its preclinical stages, such as subjective cognitive decline (SCD), is crucial for a timely management of the condition. However, current early diagnostic methods are unsuitable for preclinical screenings due to limited availability and diagnostic reliability. Additionally, reliance on invasive and scarcely available methods exacerbates the underdiagnosis of AD in its preclinical forms. We introduce an early diagnostic pipeline based on the Digital Alzheimer’s Disease Diagnosis (DADD) digital twin model, which derives personalized AD biomarkers from non-invasive electroencephalographic (EEG) recordings. These biomarkers reconstruct patient-specific neurodegeneration, capturing synaptic and connectivity degeneration mechanisms. Digital biomarkers were used to predict cerebrospinal fluid (CSF) biomarker positivity for AD and clinical conversions at follow-up in 124 participants with varying degrees of cognitive decline, including a control group of 19 healthy subjects. Digital biomarkers derived from the DADD model: i) Robustly distinguished SCD from healthy participants, improving classification accuracy by 7% compared to standard EEG biomarkers; ii) Identified patients positive for CSF biomarkers of AD with 88% accuracy (significantly outperforming standard EEG biomarkers, which achieved 58% accuracy); iii) Predicted follow-up conversions to clinical cognitive decline with 87% accuracy (compared to 54% accuracy for standard EEG biomarkers). The DADD model provided robust digital AD biomarkers with strong diagnostic and prognostic value for preclinical AD, enabling the prediction of CSF biomarkers and clinical conversions using only non-invasive EEG recordings. This is particularly important as preclinical patients, such as those with SCD, are often excluded from diagnostic procedures like lumbar puncture. Predicting CSF biomarkers by combining digital twins with non-invasive recordings could revolutionize AD diagnosis in its early stages, paving the way for the clinical application of digital twins in AD diagnostics. Clinical Trial identifier: NCT05569083 (submitted 2022–08-24).
- 31 May 2025
- Published by Springer Nature in Alzheimer's Research & Therapy
- Vol. 17 (1) , 1-16
Open AccessDigital twins and non-invasive recordings enable early diagnosis of Alzheimer’s disease
The diagnosis of Alzheimer’s disease (AD) in its preclinical stages, such as subjective cognitive decline (SCD), is crucial for a timely management of the condition. However, current early diagnostic methods are unsuitable for preclinical screenings due to limited availability and diagnostic reliability. Additionally, reliance on invasive and scarcely available methods exacerbates the underdiagnosis of AD in its preclinical forms. We introduce an early diagnostic pipeline based on the Digital Alzheimer’s Disease Diagnosis (DADD) digital twin model, which derives personalized AD biomarkers from non-invasive electroencephalographic (EEG) recordings. These biomarkers reconstruct patient-specific neurodegeneration, capturing synaptic and connectivity degeneration mechanisms. Digital biomarkers were used to predict cerebrospinal fluid (CSF) biomarker positivity for AD and clinical conversions at follow-up in 124 participants with varying degrees of cognitive decline, including a control group of 19 healthy subjects. Digital biomarkers derived from the DADD model: i) Robustly distinguished SCD from healthy participants, improving classification accuracy by 7% compared to standard EEG biomarkers; ii) Identified patients positive for CSF biomarkers of AD with 88% accuracy (significantly outperforming standard EEG biomarkers, which achieved 58% accuracy); iii) Predicted follow-up conversions to clinical cognitive decline with 87% accuracy (compared to 54% accuracy for standard EEG biomarkers). The DADD model provided robust digital AD biomarkers with strong diagnostic and prognostic value for preclinical AD, enabling the prediction of CSF biomarkers and clinical conversions using only non-invasive EEG recordings. This is particularly important as preclinical patients, such as those with SCD, are often excluded from diagnostic procedures like lumbar puncture. Predicting CSF biomarkers by combining digital twins with non-invasive recordings could revolutionize AD diagnosis in its early stages, paving the way for the clinical application of digital twins in AD diagnostics. Clinical Trial identifier: NCT05569083 (submitted 2022–08-24).
- Open Access
- 1 January 2024
- Published by Wiley in Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring
- Vol. 16 (1) , e12526
Personalized modeling of Alzheimer's disease progression estimates neurodegeneration severity from EEG recordings
INTRODUCTION: Early identification of Alzheimer's disease (AD) is necessary for a timely onset of therapeutic care. However, cortical structural alterations associated with AD are difficult to discern.METHODS: We developed a cortical model of AD‐related neurodegeneration accounting for slowing of local dynamics and global connectivity degradation. In a monocentric study we collected electroencephalography (EEG) recordings at rest from participants in healthy (HC, n = 17), subjective cognitive decline (SCD, n = 58), and mild cognitive impairment (MCI, n = 44) conditions. For each patient, we estimated neurodegeneration model parameters based on individual EEG recordings.RESULTS: Our model outperformed standard EEG analysis not only in discriminating between HC and MCI conditions (F1 score 0.95 vs 0.75) but also in identifying SCD patients with biological hallmarks of AD in the cerebrospinal fluid (recall 0.87 vs 0.50).DISCUSSION: Personalized models could (1) support classification of MCI, (2) assess the presence of AD pathology, and (3) estimate the risk of cognitive decline progression, based only on economical and non‐invasive EEG recordings.Highlights: Personalized cortical model estimating structural alterations from EEG recordings. Discrimination of Mild Cognitive Impairment (MCI) and Healthy (HC) subjects (95%) Prediction of biological markers of Alzheimer's in Subjective Decline (SCD) Subjects (87%) Transition correctly predicted for 3/3 subjects that converted from SCD to MCI after 1y
- 1 January 2024
- Published by Wiley in Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring
- Vol. 16 (1) , e12526
Open AccessPersonalized modeling of Alzheimer's disease progression estimates neurodegeneration severity from EEG recordings
INTRODUCTION: Early identification of Alzheimer's disease (AD) is necessary for a timely onset of therapeutic care. However, cortical structural alterations associated with AD are difficult to discern.METHODS: We developed a cortical model of AD‐related neurodegeneration accounting for slowing of local dynamics and global connectivity degradation. In a monocentric study we collected electroencephalography (EEG) recordings at rest from participants in healthy (HC, n = 17), subjective cognitive decline (SCD, n = 58), and mild cognitive impairment (MCI, n = 44) conditions. For each patient, we estimated neurodegeneration model parameters based on individual EEG recordings.RESULTS: Our model outperformed standard EEG analysis not only in discriminating between HC and MCI conditions (F1 score 0.95 vs 0.75) but also in identifying SCD patients with biological hallmarks of AD in the cerebrospinal fluid (recall 0.87 vs 0.50).DISCUSSION: Personalized models could (1) support classification of MCI, (2) assess the presence of AD pathology, and (3) estimate the risk of cognitive decline progression, based only on economical and non‐invasive EEG recordings.Highlights: Personalized cortical model estimating structural alterations from EEG recordings. Discrimination of Mild Cognitive Impairment (MCI) and Healthy (HC) subjects (95%) Prediction of biological markers of Alzheimer's in Subjective Decline (SCD) Subjects (87%) Transition correctly predicted for 3/3 subjects that converted from SCD to MCI after 1y
- Open Access
Personalized modeling of neurodegeneration determines dementia severity from EEG recordings
INTRODUCTION: Early identification of dementia is necessary for a timely onset of therapeutic care. However, cortical structural alterations associated with early dementia are difficult to disclose. METHODS: We developed a cortical model of dementia-related neurodegeneration accounting for slowing of local dynamics and global connectivity degradation. We collected EEG recordings at rest from subjects in healthy (HC), Subjective Cognitive Decline (SCD), and Mild Cognitive Impairment (MCI) condition. For each patient, we estimated neurodegeneration model parameters based on individual EEG recordings. RESULTS: Our model outperformed standard EEG analysis not only in discriminating between HC and MCI conditions (F1 score 0.95 vs 0.85) but also in identifying SCD patients with biological hallmarks of Alzheimer’s disease in the cerebrospinal fluid (recall 0.87 vs 0.50). DISCUSSION: Personalized neurodegeneration models could both support classification of MCI and assess the risk of progression from SCD to Alzheimer based only on economical and non-invasive EEG recording ClinicalTrials.gov Identifier: NCT05569083
Open AccessPersonalized modeling of neurodegeneration determines dementia severity from EEG recordings
INTRODUCTION: Early identification of dementia is necessary for a timely onset of therapeutic care. However, cortical structural alterations associated with early dementia are difficult to disclose. METHODS: We developed a cortical model of dementia-related neurodegeneration accounting for slowing of local dynamics and global connectivity degradation. We collected EEG recordings at rest from subjects in healthy (HC), Subjective Cognitive Decline (SCD), and Mild Cognitive Impairment (MCI) condition. For each patient, we estimated neurodegeneration model parameters based on individual EEG recordings. RESULTS: Our model outperformed standard EEG analysis not only in discriminating between HC and MCI conditions (F1 score 0.95 vs 0.85) but also in identifying SCD patients with biological hallmarks of Alzheimer’s disease in the cerebrospinal fluid (recall 0.87 vs 0.50). DISCUSSION: Personalized neurodegeneration models could both support classification of MCI and assess the risk of progression from SCD to Alzheimer based only on economical and non-invasive EEG recording ClinicalTrials.gov Identifier: NCT05569083