Predicting and Characterizing Neurodegenerative Subtypes with Multimodal Neurocognitive Signatures of Social and Cognitive Processes

A systematic review of cognitive and behavioral tools to differentiate behavioral variant frontotemporal dementia from other conditions

The behavioral variant of frontotemporal dementia (bvFTD) is thought to be the commonest clinical presentation of frontotemporal lobar degeneration and is predominantly characterized by changes in behavior. In patients lacking unequivocal biomarker evidence of frontotemporal neurodegeneration, the clinical diagnosis of bvFTD is often unstable. In response, we conducted a systematic review and critical appraisal of cognitive and behavioral tools that have sought to differentiate bvFTD from other conditions. A systematic literature review of PubMed, Scopus, and Web of Science was conducted on December 31, 2023 for cognitive and behavioral tools that differentiated bvFTD from other cohorts. Ninety-six studies were included. The quality appraisal of almost all studies was low and introduced a high risk of bias. The few studies that were of high quality had a prospective study design and recruited patients suspected (but not yet confirmed) to have bvFTD. These studies reported that behavioral tools (e.g., the Frontal Behavioral Inventory) and social cognition tests (e.g., the Ekman's Faces Test) had good test performance in differentiating bvFTD from a broad range of psychiatric and neurological conditions. Importantly, the review highlighted the extreme paucity of studies that have evaluated methods where, in Bayesian terms, there is genuine clinical uncertainty regarding a diagnosis of bvFTD. Most studies used healthy controls of typical Alzheimer's disease as comparators-groups that often have negligible pretest probability of bvFTD. In response, we propose a study design checklist for studies seeking to develop diagnostic algorithms in bvFTD research.

Self-supervised multimodal learning for group inferences from MRI data: Discovering disorder-relevant brain regions and multimodal links

In recent years, deep learning approaches have gained significant attention in predicting brain disorders using neuroimaging data. However, conventional methods often rely on single-modality data and supervised models, which provide only a limited perspective of the intricacies of the highly complex brain. Moreover, the scarcity of accurate diagnostic labels in clinical settings hinders the applicability of the supervised models. To address these limitations, we propose a novel self-supervised framework for extracting multiple representations from multimodal neuroimaging data to enhance group inferences and enable analysis without resorting to labeled data during pre-training. Our approach leverages Deep InfoMax (DIM), a self-supervised methodology renowned for its efficacy in learning representations by estimating mutual information without the need for explicit labels. While DIM has shown promise in predicting brain disorders from single-modality MRI data, its potential for multimodal data remains untapped. This work extends DIM to multimodal neuroimaging data, allowing us to identify disorder-relevant brain regions and explore multimodal links. We present compelling evidence of the efficacy of our multimodal DIM analysis in uncovering disorder-relevant brain regions, including the hippocampus, caudate, insula, - and multimodal links with the thalamus, precuneus, and subthalamus hypothalamus. Our self-supervised representations demonstrate promising capabilities in predicting the presence of brain disorders across a spectrum of Alzheimer's phenotypes. Comparative evaluations against state-of-the-art unsupervised methods based on autoencoders, canonical correlation analysis, and supervised models highlight the superiority of our proposed method in achieving improved classification performance, capturing joint information, and interpretability capabilities. The computational efficiency of the decoder-free strategy enhances its practical utility, as it saves compute resources without compromising performance. This work offers a significant step forward in addressing the challenge of understanding multimodal links in complex brain disorders, with potential applications in neuroimaging research and clinical diagnosis.

The BrainLat project, a multimodal neuroimaging dataset of neurodegeneration from underrepresented backgrounds

The Latin American Brain Health Institute (BrainLat) has released a unique multimodal neuroimaging dataset of 780 participants from Latin American. The dataset includes 530 patients with neurodegenerative diseases such as Alzheimer's disease (AD), behavioral variant frontotemporal dementia (bvFTD), multiple sclerosis (MS), Parkinson's disease (PD), and 250 healthy controls (HCs). This dataset (62.7 ± 9.5 years, age range 21-89 years) was collected through a multicentric effort across five Latin American countries to address the need for affordable, scalable, and available biomarkers in regions with larger inequities. The BrainLat is the first regional collection of clinical and cognitive assessments, anatomical magnetic resonance imaging (MRI), resting-state functional MRI (fMRI), diffusion-weighted MRI (DWI), and high density resting-state electroencephalography (EEG) in dementia patients. In addition, it includes demographic information about harmonized recruitment and assessment protocols. The dataset is publicly available to encourage further research and development of tools and health applications for neurodegeneration based on multimodal neuroimaging, promoting the assessment of regional variability and inclusion of underrepresented participants in research.

The neurocognitive impact of loneliness and social networks on social adaptation

Social adaptation arises from the interaction between the individual and the social environment. However, little empirical evidence exists regarding the relationship between social contact and social adaptation. We propose that loneliness and social networks are key factors explaining social adaptation. Sixty-four healthy subjects with no history of psychiatric conditions participated in this study. All participants completed self-report questionnaires about loneliness, social network, and social adaptation. On a separate day, subjects underwent a resting state fMRI recording session. A hierarchical regression model on self-report data revealed that loneliness and social network were negatively and positively associated with social adaptation. Functional connectivity (FC) analysis showed that loneliness was associated with decreased FC between the fronto-amygdalar and fronto-parietal regions. In contrast, the social network was positively associated with FC between the fronto-temporo-parietal network. Finally, an integrative path model examined the combined effects of behavioral and brain predictors of social adaptation. The model revealed that social networks mediated the effects of loneliness on social adaptation. Further, loneliness-related abnormal brain FC (previously shown to be associated with difficulties in cognitive control, emotion regulation, and sociocognitive processes) emerged as the strongest predictor of poor social adaptation. Findings offer insights into the brain indicators of social adaptation and highlight the role of social networks as a buffer against the maladaptive effects of loneliness. These findings can inform interventions aimed at minimizing loneliness and promoting social adaptation and are especially relevant due to the high prevalence of loneliness around the globe. These findings also serve the study of social adaptation since they provide potential neurocognitive factors that could influence social adaptation.

The impact of regional heterogeneity in whole-brain dynamics in the presence of oscillations

Large variability exists across brain regions in health and disease, considering their cellular and molecular composition, connectivity, and function. Large-scale whole-brain models comprising coupled brain regions provide insights into the underlying dynamics that shape complex patterns of spontaneous brain activity. In particular, biophysically grounded mean-field whole-brain models in the asynchronous regime were used to demonstrate the dynamical consequences of including regional variability. Nevertheless, the role of heterogeneities when brain dynamics are supported by synchronous oscillating state, which is a ubiquitous phenomenon in brain, remains poorly understood. Here, we implemented two models capable of presenting oscillatory behavior with different levels of abstraction: a phenomenological Stuart-Landau model and an exact mean-field model. The fit of these models informed by structural- to functional-weighted MRI signal (T1w/T2w) allowed us to explore the implication of the inclusion of heterogeneities for modeling resting-state fMRI recordings from healthy participants. We found that disease-specific regional functional heterogeneity imposed dynamical consequences within the oscillatory regime in fMRI recordings from neurodegeneration with specific impacts on brain atrophy/structure (Alzheimer's patients). Overall, we found that models with oscillations perform better when structural and functional regional heterogeneities are considered, showing that phenomenological and biophysical models behave similarly at the brink of the Hopf bifurcation.

Heterogeneous factors influence social cognition across diverse settings in brain health and age-related diseases

Aging may diminish social cognition, which is crucial for interaction with others, and significant changes in this capacity can indicate pathological processes like dementia. However, the extent to which non-specific factors explain variability in social cognition performance, especially among older adults and in global settings, remains unknown. A computational approach assessed combined heterogeneous contributors to social cognition in a diverse sample of 1063 older adults from 9 countries. Support vector regressions predicted the performance in emotion recognition, mentalizing, and a total social cognition score from a combination of disparate factors, including clinical diagnosis (healthy controls, subjective cognitive complaints, mild cognitive impairment, Alzheimer's disease, behavioral variant frontotemporal dementia), demographics (sex, age, education, and country income as a proxy of socioeconomic status), cognition (cognitive and executive functions), structural brain reserve, and in-scanner motion artifacts. Cognitive and executive functions and educational level consistently emerged among the top predictors of social cognition across models. Such non-specific factors showed more substantial influence than diagnosis (dementia or cognitive decline) and brain reserve. Notably, age did not make a significant contribution when considering all predictors. While fMRI brain networks did not show predictive value, head movements significantly contributed to emotion recognition. Models explained between 28-44% of the variance in social cognition performance. Results challenge traditional interpretations of age-related decline, patient-control differences, and brain signatures of social cognition, emphasizing the role of heterogeneous factors. Findings advance our understanding of social cognition in brain health and disease, with implications for predictive models, assessments, and interventions.

Social concepts and the cerebellum: behavioural and functional connectivity signatures in cerebellar ataxic patients

Neurocognitive research on social concepts underscores their reliance on fronto-temporo-limbic regions mediating broad socio-cognitive skills. Yet, the field has neglected another structure increasingly implicated in social cognition: the cerebellum. The present exploratory study examines this link combining a novel naturalistic text paradigm, a relevant atrophy model and functional magnetic resonance imaging. Fifteen cerebellar ataxia (CA) patients with focal cerebellar atrophy and 29 matched controls listened to a social text (highlighting interpersonal events) as well as a non-social text (focused on a single person's actions), and answered comprehension questionnaires. We compared behavioural outcomes between groups and examined their association with cerebellar connectivity. CA patients showed deficits in social text comprehension and normal scores in the non-social text. Also, social text outcomes in controls selectively correlated with connectivity between the cerebellum and key regions subserving multi-modal semantics and social cognition, including the superior and medial temporal gyri, the temporal pole and the insula. Conversely, brain-behaviour associations involving the cerebellum were abolished in the patients. Thus, cerebellar structures and connections seem involved in processing social concepts evoked by naturalistic discourse. Such findings invite new theoretical and translational developments integrating social neuroscience with embodied semantics. This article is part of the theme issue 'Concepts in interaction: social engagement and inner experiences'.

Emotional descriptions increase accidental harm punishment and its cortico-limbic signatures during moral judgment in autism

Individuals with autism spectrum disorder (ASD) present difficulties in integrating mental state information in complex moral tasks. Yet, ASD research has not examined whether this process is influenced by emotions, let alone while capturing its neural bases. We investigated how language-induced emotions modulate intent-based moral judgment in ASD. In a fMRI task, 30 adults with ASD and 27 neurotypical controls read vignettes whose protagonists commit harm either accidentally or intentionally, and then decided how much punishment the protagonist deserved. Emotional content was manipulated across scenarios through the use of graphic language (designed to trigger arousing negative responses) vs. plain (just-the-facts, emotionless) language. Off-line functional connectivity correlates of task performance were also analyzed. In ASD, emotional (graphic) descriptions amplified punishment ratings of accidental harms, associated with increased activity in fronto-temporo-limbic, precentral, and postcentral/supramarginal regions (critical for emotional and empathic processes), and reduced connectivity among the orbitofrontal cortex and the angular gyrus (involved in mentalizing). Language manipulation did not influence intentional harm processing in ASD. In conclusion, in arousing and ambiguous social situations that lack intentionality clues (i.e. graphic accidental harm scenarios), individuals with ASD would misuse their emotional responses as the main source of information to guide their moral decisions. Conversely, in face of explicit harmful intentions, they would be able to compensate their socioemotional alterations and assign punishment through non-emotional pathways. Despite limitations, such as the small sample size and low ecological validity of the task, results of the present study proved reliable and have relevant theoretical and translational implications.

Machine learning for predicting neurodegenerative diseases in the general older population: a cohort study

BACKGROUND

In the older general population, neurodegenerative diseases (NDs) are associated with increased disability, decreased physical and cognitive function. Detecting risk factors can help implement prevention measures. Using deep neural networks (DNNs), a machine-learning algorithm could be an alternative to Cox regression in tabular datasets with many predictive features. We aimed to compare the performance of different types of DNNs with regularized Cox proportional hazards models to predict NDs in the older general population.

METHODS

We performed a longitudinal analysis with participants of the English Longitudinal Study of Ageing. We included men and women with no NDs at baseline, aged 60 years and older, assessed every 2 years from 2004 to 2005 (wave2) to 2016-2017 (wave 8). The features were a set of 91 epidemiological and clinical baseline variables. The outcome was new events of Parkinson's, Alzheimer or dementia. After applying multiple imputations, we trained three DNN algorithms: Feedforward, TabTransformer, and Dense Convolutional (Densenet). In addition, we trained two algorithms based on Cox models: Elastic Net regularization (CoxEn) and selected features (CoxSf).

RESULTS

5433 participants were included in wave 2. During follow-up, 12.7% participants developed NDs. Although the five models predicted NDs events, the discriminative ability was superior using TabTransformer (Uno's C-statistic (coefficient (95% confidence intervals)) 0.757 (0.702, 0.805). TabTransformer showed superior time-dependent balanced accuracy (0.834 (0.779, 0.889)) and specificity (0.855 (0.0.773, 0.909)) than the other models. With the CoxSf (hazard ratio (95% confidence intervals)), age (10.0 (6.9, 14.7)), poor hearing (1.3 (1.1, 1.5)) and weight loss 1.3 (1.1, 1.6)) were associated with a higher DNN risk. In contrast, executive function (0.3 (0.2, 0.6)), memory (0, 0, 0.1)), increased gait speed (0.2, (0.1, 0.4)), vigorous physical activity (0.7, 0.6, 0.9)) and higher BMI (0.4 (0.2, 0.8)) were associated with a lower DNN risk.

CONCLUSION

TabTransformer is promising for prediction of NDs with heterogeneous tabular datasets with numerous features. Moreover, it can handle censored data. However, Cox models perform well and are easier to interpret than DNNs. Therefore, they are still a good choice for NDs.

Moral Emotions and Their Brain Structural Correlates Across Neurodegenerative Disorders

BACKGROUND

Although social cognition is compromised in patients with neurodegenerative disorders such as behavioral variant frontotemporal dementia (bvFTD) and Alzheimer's disease (AD), research on moral emotions and their neural correlates in these populations is scarce. No previous study has explored the utility of moral emotions, compared to and in combination with classical general cognitive state tools, to discriminate bvFTD from AD patients.

OBJECTIVE

To examine self-conscious (guilt and embarrassment) and other-oriented (pity and indignation) moral emotions, their subjective experience, and their structural brain underpinnings in bvFTD (n = 31) and AD (n = 30) patients, compared to healthy controls (n = 37). We also explored the potential utility of moral emotions measures to discriminate bvFTD from AD.

METHODS

We used a modified version of the Moral Sentiment Task measuring the participants' accuracy scores and their emotional subjective experiences.

RESULTS

bvFTD patients exhibited greater impairments in self-conscious and other-oriented moral emotions as compared with AD patients and healthy controls. Moral emotions combined with general cognitive state tools emerged as useful measures to discriminate bvFTD from AD patients. In bvFTD patients, lower moral emotions scores were associated with lower gray matter volumes in caudate nucleus and inferior and middle temporal gyri. In AD, these scores were associated with lower gray matter volumes in superior and middle frontal gyri, middle temporal gyrus, inferior parietal lobule and supramarginal gyrus.

CONCLUSION

These findings contribute to a better understanding of moral emotion deficits across neurodegenerative disorders, highlighting the potential benefits of integrating this domain into the clinical assessment.

Multiclass characterization of frontotemporal dementia variants via multimodal brain network computational inference

Characterizing a particular neurodegenerative condition against others possible diseases remains a challenge along clinical, biomarker, and neuroscientific levels. This is the particular case of frontotemporal dementia (FTD) variants, where their specific characterization requires high levels of expertise and multidisciplinary teams to subtly distinguish among similar physiopathological processes. Here, we used a computational approach of multimodal brain networks to address simultaneous multiclass classification of 298 subjects (one group against all others), including five FTD variants: behavioral variant FTD, corticobasal syndrome, nonfluent variant primary progressive aphasia, progressive supranuclear palsy, and semantic variant primary progressive aphasia, with healthy controls. Fourteen machine learning classifiers were trained with functional and structural connectivity metrics calculated through different methods. Due to the large number of variables, dimensionality was reduced, employing statistical comparisons and progressive elimination to assess feature stability under nested cross-validation. The machine learning performance was measured through the area under the receiver operating characteristic curves, reaching 0.81 on average, with a standard deviation of 0.09. Furthermore, the contributions of demographic and cognitive data were also assessed via multifeatured classifiers. An accurate simultaneous multiclass classification of each FTD variant against other variants and controls was obtained based on the selection of an optimum set of features. The classifiers incorporating the brain's network and cognitive assessment increased performance metrics. Multimodal classifiers evidenced specific variants' compromise, across modalities and methods through feature importance analysis. If replicated and validated, this approach may help to support clinical decision tools aimed to detect specific affectations in the context of overlapping diseases.

Multimodal Neurocognitive Markers of Naturalistic Discourse Typify Diverse Neurodegenerative Diseases

Neurodegeneration has multiscalar impacts, including behavioral, neuroanatomical, and neurofunctional disruptions. Can disease-differential alterations be captured across such dimensions using naturalistic stimuli? To address this question, we assessed comprehension of four naturalistic stories, highlighting action, nonaction, social, and nonsocial events, in Parkinson's disease (PD) and behavioral variant frontotemporal dementia (bvFTD) relative to Alzheimer's disease patients and healthy controls. Text-specific correlates were evaluated via voxel-based morphometry, spatial (fMRI), and temporal (hd-EEG) functional connectivity. PD patients presented action-text deficits related to the volume of action-observation regions, connectivity across motor-related and multimodal-semantic hubs, and frontal hd-EEG hypoconnectivity. BvFTD patients exhibited social-text deficits, associated with atrophy and spatial connectivity patterns along social-network hubs, alongside right frontotemporal hd-EEG hypoconnectivity. Alzheimer's disease patients showed impairments in all stories, widespread atrophy and spatial connectivity patterns, and heightened occipitotemporal hd-EEG connectivity. Our framework revealed disease-specific signatures across behavioral, neuroanatomical, and neurofunctional dimensions, highlighting the sensitivity and specificity of a single naturalistic task. This investigation opens a translational agenda combining ecological approaches and multimodal cognitive neuroscience for the study of neurodegeneration.

Allostatic-Interoceptive Overload in Frontotemporal Dementia

BACKGROUND

The predictive coding theory of allostatic-interoceptive load states that brain networks mediating autonomic regulation and interoceptive-exteroceptive balance regulate the internal milieu to anticipate future needs and environmental demands. These functions seem to be distinctly compromised in behavioral variant frontotemporal dementia (bvFTD), including alterations of the allostatic-interoceptive network (AIN). Here, we hypothesize that bvFTD is typified by an allostatic-interoceptive overload.

METHODS

We assessed resting-state heartbeat evoked potential (rsHEP) modulation as well as its behavioral and multimodal neuroimaging correlates in patients with bvFTD relative to healthy control subjects and patients with Alzheimer's disease (N = 94). We measured 1) resting-state electroencephalography (to assess the rsHEP, prompted by visceral inputs and modulated by internal body sensing), 2) associations between rsHEP and its neural generators (source location), 3) cognitive disturbances (cognitive state, executive functions, facial emotion recognition), 4) brain atrophy, and 5) resting-state functional magnetic resonance imaging functional connectivity (AIN vs. control networks).

RESULTS

Relative to healthy control subjects and patients with Alzheimer's disease, patients with bvFTD presented more negative rsHEP amplitudes with sources in critical hubs of the AIN (insula, amygdala, somatosensory cortex, hippocampus, anterior cingulate cortex). This exacerbated rsHEP modulation selectively predicted the patients' cognitive profile (including cognitive decline, executive dysfunction, and emotional impairments). In addition, increased rsHEP modulation in bvFTD was associated with decreased brain volume and connectivity of the AIN. Machine learning results confirmed AIN specificity in predicting the bvFTD group.

CONCLUSIONS

Altogether, these results suggest that bvFTD may be characterized by an allostatic-interoceptive overload manifested in ongoing electrophysiological markers, brain atrophy, functional networks, and cognition.

Multidimensional Clinical Assessment in Frontotemporal Dementia and Its Spectrum in Latin America and the Caribbean: A Narrative Review and a Glance at Future Challenges

Frontotemporal dementia (FTD) is the third most common form of dementia across all age groups and is a leading cause of early-onset dementia. The Frontotemporal dementia (FTD) includes a spectrum of diseases that are classified according to their clinical presentation and patterns of neurodegeneration. There are two main types of FTD: behavioral FTD variant (bvFTD), characterized by a deterioration in social function, behavior, and personality; and primary progressive aphasias (PPA), characterized by a deficit in language skills. There are other types of FTD-related disorders that present motor impairment and/or parkinsonism, including FTD with motor neuron disease (FTD-MND), progressive supranuclear palsy (PSP), and corticobasal syndrome (CBS). The FTD and its associated disorders present great clinical heterogeneity. The diagnosis of FTD is based on the identification through clinical assessments of a specific clinical phenotype of impairments in different domains, complemented by an evaluation through instruments, i.e., tests and questionnaires, validated for the population under study, thus, achieving timely detection and treatment. While the prevalence of dementia in Latin America and the Caribbean (LAC) is increasing rapidly, there is still a lack of standardized instruments and consensus for FTD diagnosis. In this context, it is important to review the published tests and questionnaires adapted and/or validated in LAC for the assessment of cognition, behavior, functionality, and gait in FTD and its spectrum. Therefore, our paper has three main goals. First, to present a narrative review of the main tests and questionnaires published in LAC for the assessment of FTD and its spectrum in six dimensions: (i) Cognitive screening; (ii) Neuropsychological assessment divided by cognitive domain; (iii) Gait assessment; (iv) Behavioral and neuropsychiatric symptoms; (v) Functional assessment; and (vi) Global Rating Scale. Second, to propose a multidimensional clinical assessment of FTD in LAC identifying the main gaps. Lastly, it is proposed to create a LAC consortium that will discuss strategies to address the current challenges in the field.

Multimodal mechanisms of human socially reinforced learning across neurodegenerative diseases

Social feedback can selectively enhance learning in diverse domains. Relevant neurocognitive mechanisms have been studied mainly in healthy persons, yielding correlational findings. Neurodegenerative lesion models, coupled with multimodal brain measures, can complement standard approaches by revealing direct multidimensional correlates of the phenomenon.

To this end, we assessed socially reinforced and non-socially reinforced learning in 40 healthy participants as well as persons with behavioural variant frontotemporal dementia (n = 21), Parkinson’s disease (n = 31) and Alzheimer’s disease (n = 20). These conditions are typified by predominant deficits in social cognition, feedback-based learning and associative learning, respectively, although all three domains may be partly compromised in the other conditions. We combined a validated behavioural task with ongoing EEG signatures of implicit learning (medial frontal negativity) and offline MRI measures (voxel-based morphometry).

In healthy participants, learning was facilitated by social feedback relative to non-social feedback. In comparison with controls, this effect was specifically impaired in behavioural variant frontotemporal dementia and Parkinson’s disease, while unspecific learning deficits (across social and non-social conditions) were observed in Alzheimer’s disease. EEG results showed increased medial frontal negativity in healthy controls during social feedback and learning. Such a modulation was selectively disrupted in behavioural variant frontotemporal dementia. Neuroanatomical results revealed extended temporo-parietal and fronto-limbic correlates of socially reinforced learning, with specific temporo-parietal associations in behavioural variant frontotemporal dementia and predominantly fronto-limbic regions in Alzheimer’s disease. In contrast, non-socially reinforced learning was consistently linked to medial temporal/hippocampal regions. No associations with cortical volume were found in Parkinson’s disease. Results are consistent with core social deficits in behavioural variant frontotemporal dementia, subtle disruptions in ongoing feedback-mechanisms and social processes in Parkinson’s disease and generalized learning alterations in Alzheimer’s disease. This multimodal approach highlights the impact of different neurodegenerative profiles on learning and social feedback.

Our findings inform a promising theoretical and clinical agenda in the fields of social learning, socially reinforced learning and neurodegeneration.