Lower functional hippocampal redundancy in mild cognitive impairment

Langella, Stephanie; Sadiq, Muhammad Usman; Mucha, Peter J.; Giovanello, Kelly S.; Dayan, Eran; Initiative, Alzheimer’s Disease Neuroimaging

doi:10.1038/s41398-020-01166-w

Cited by 29 publications

(52 citation statements)

References 45 publications

(37 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These findings are consistent with earlier studies on AD pathology, where atrophy was reported in the hippocampus, amygdala, and entorhinal cortex. 26 , 53 , 54 These findings are also consistent with recent findings on the possible neuroprotective role of redundancy in the hippocampus 55 and other large-scale brain networks, 56 which may contribute to the observed differences between the subgroups. Our results also suggest that the occipital lobe played a more minor role in the performance of the model, consistent with Braak’s staging scheme, 57 where the occipital lobe is shown to be affected only at later stages of AD.…”

Section: Discussionsupporting

confidence: 90%

Subtyping of mild cognitive impairment using a deep learning model based on brain atrophy patterns

Kwak

Giovanello

Bozoki

et al. 2021

Cell Reports Medicine

Self Cite

View full text Add to dashboard Cite

Summary Trajectories of cognitive decline vary considerably among individuals with mild cognitive impairment (MCI). To address this heterogeneity, subtyping approaches have been developed, with the objective of identifying more homogeneous subgroups. To date, subtyping of MCI has been based primarily on cognitive measures, often resulting in indistinct boundaries between subgroups and limited validity. Here, we introduce a subtyping method for MCI based solely upon brain atrophy. We train a deep learning model to differentiate between Alzheimer’s disease (AD) and cognitively normal (CN) subjects based on whole-brain MRI features. We then deploy the trained model to classify MCI subjects based on whole-brain gray matter resemblance to AD-like or CN-like patterns. We subsequently validate the subtyping approach using cognitive, clinical, fluid biomarker, and molecular imaging data. Overall, the results suggest that atrophy patterns in MCI are sufficiently heterogeneous and can thus be used to subtype individuals into biologically and clinically meaningful subgroups.

show abstract

Section: Discussionsupporting

confidence: 90%

Subtyping of mild cognitive impairment using a deep learning model based on brain atrophy patterns

Kwak

Giovanello

Bozoki

et al. 2021

Cell Reports Medicine

Self Cite

View full text Add to dashboard Cite

show abstract

“…As in previous reports with neurodegenerative diseases, 30 , 32 , 104–106 because our hypotheses hinged on differences between each neurodegenerative group and healthy controls, and given that demographic and behavioural features were not matched across neurodegenerative samples (bvFTD versus Parkinson’s disease versus Alzheimer’s disease), we focused on pairwise comparisons between demographically matched tandems: healthy controls versus bvFTD, healthy controls versus Parkinson’s disease, healthy controls versus Alzheimer’s disease ( Table 2 ). In addition, given that a significant difference was found in sex between bvFTD and healthy controls, we conducted additional group comparison analyses of covariance using permutation testing controlling for sex 107 ( Supplementary material ). Moreover, to rule out potential confounds of facial emotion recognition disturbances in bvFTD (particularly, for negative emotions), 108, 109 we also conducted additional group comparison analyses of covariance using permutation testing and controlling for feedback valence recognition ( Supplementary material ).…”

Section: Methodsmentioning

confidence: 99%

Multimodal mechanisms of human socially reinforced learning across neurodegenerative diseases

Legaz

Abrevaya

Dottori

et al. 2021

Brain

View full text Add to dashboard Cite

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 behavioral 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 behavioral task with ongoing electroencephalographic (EEG) signatures of implicit learning (medial frontal negativity) and offline magnetic resonance imaging (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 behavioral 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 behavioral variant frontotemporal dementia. Neuroanatomical results revealed extended temporo-parietal and fronto-limbic correlates of socially reinforced learning, with specific temporo-parietal associations in behavioral 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 behavioral 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.

show abstract

“…There has been growing use of ML to determine if brain network metrics can serve as classifiers of brain disorders with several high-profile reviews recently published ( Bassett et al, 2020 ; Braun et al, 2018 ; Parkes et al, 2020 ; Vu et al, 2018 ). Many of the canonical networks identified in rsfMRI studies (e.g., default mode network) have been of critical focus in studies of large-scale network plasticity in a range of brain disorders including schizophrenia ( de Filippis et al, 2019 ; Lefort-Besnard et al, 2018 ; Progar & May, 1988 ; Steardo et al, 2020 ), autism ( L. Chen et al, 2020 ; Glerean et al, 2016 ; Hegarty et al, 2017 ), Alzheimer’s disease and related dementias ( Langella et al, 2021 ; Pellegrini et al, 2018 ; Salvatore et al, 2015 ), and brain injury ( Bonnelle et al, 2012 ; Caeyenberghs et al, 2017 ; Gilbert et al, 2018 ; Roy et al, 2017 ).…”

Section: Introductionmentioning

confidence: 99%

Feeding the machine: Challenges to reproducible predictive modeling in resting-state connectomics

Cwiek

Rajtmajer

Wyble

et al. 2022

Network Neuroscience

View full text Add to dashboard Cite

In this critical review, we examine the application of predictive models, e.g. classifiers, trained using Machine Learning (ML) to assist in interpretation of functional neuroimaging data. Our primary goal is to summarize how ML is being applied and critically assess common practices. Our review covers 250 studies published using ML and resting-state functional MRI (fMRI) to infer various dimensions of the human functional connectome. Results for hold-out (“lockbox”) performance was, on average, ~13% less accurate than performance measured through cross-validation alone, highlighting the importance of lockbox data which was included in only 16% of the studies. There was also a concerning lack of transparency across the key steps in training and evaluating predictive models. The summary of this literature underscores the importance of the use of a lockbox and highlights several methodological pitfalls that can be addressed by the imaging community. We argue that, ideally, studies are motivated both by the reproducibility and generalizability of findings as well as the potential clinical significance of the insights. We offer recommendations for principled integration of machine learning into the clinical neurosciences with the goal of advancing imaging biomarkers of brain disorders, understanding causative determinants for health risks, and parsing heterogeneous patient outcomes.

show abstract

Lower functional hippocampal redundancy in mild cognitive impairment

Cited by 29 publications

References 45 publications

Subtyping of mild cognitive impairment using a deep learning model based on brain atrophy patterns

Subtyping of mild cognitive impairment using a deep learning model based on brain atrophy patterns

Multimodal mechanisms of human socially reinforced learning across neurodegenerative diseases

Feeding the machine: Challenges to reproducible predictive modeling in resting-state connectomics

Contact Info

Product

Resources

About