A robust brain signature region approach for episodic memory performance in older adults

Jolly, Amy; Hampshire, Adam

doi:10.1093/brain/awab140

Cited by 4 publications

(11 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They constitute a form of "data reduction" that is valuable for constructing tractable models. On the other hand, atlas regions do not necessarily align with networked locations that underlie behavioral outcomes of interest (Jolly and Hampshire, 2021) and this may explain why their explanatory performance falls short of that attained by signature models, seen in this work (Fig. 6) and our previous study.…”

Section: Relations To Brain Atlases and Theory-driven Modelscontrasting

confidence: 63%

“…Though it incorporated the four atlas regions most heavily overlapped by our signature masks, its fit performance was still lower than the signatures except in one case. Thus, one of the advantages of the exploratory signature approach may be that its selection process implicitly accounts for connections of disparate regions communicating with each other in networks associated with behavioral outcomes (Genon et al, 2018; Jolly and Hampshire, 2021). This may contribute to its superior explanatory power over that of pre-selected regions.…”

Section: Discussionmentioning

confidence: 99%

“…An important strength of the signature approach is that it proposes hypothesis-free, exploratory computation of brain regional measures maximally associated to outcome (Bakkour et al, 2009; Dickerson et al, 2009; Fletcher et al, 2021b; Jolly and Hampshire, 2021). However, achieving this promise necessarily incurs substantial difficulties (Masouleh et al, 2019).…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Brain signatures for neuropsychological and everyday memory achieve high replicability and explanatory power in two data cohorts

Fletcher

Farias

DeCarli

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: The brain signature of cognition concept has garnered interest as a data-driven, exploratory approach to better understand key brain regions involved in specific cognitive functions, with the potential to maximally characterize brain substrates of clinical outcomes. However, to be a robust brain phenotype, the signature approach requires a statistical foundation showing that model performance replicates across a variety of cohorts. Here, we outline a procedure that provides this foundation for a signature models of two memory-related behavioral domains. Method: In each of two independent data cohorts, we derived regional brain gray matter thickness associations for neuropsychological and everyday cognition memory, testing for replicability. In each cohort we computed regional association to outcome in 40 randomly selected discovery subsets of size 400; we generated spatial overlap frequency maps and selected high-frequency regions as consensus signature masks for each cohort. We tested replicability by comparing cohort-based consensus model fits in all discovery sets. We tested explanatory power in each full cohort, compare signature model fits with competing standard models of each outcome. Result: Spatial replications produced strongly convergent consensus signature regions derived from UCD and ADNI. Consensus model fits were highly correlated in 40 random subsets of each cohort indicating high replicability. In comparisons over each full cohort, signature models outperformed other models with one exception. Conclusion: Multiple random model generations, followed by consensus selection of regional brain substrates, produced signature models that replicated model fits to outcome and outperformed other commonly used measures. Robust biomarkers of cognition and everyday function may be achievable by this method.

show abstract

Section: Relations To Brain Atlases and Theory-driven Modelscontrasting

confidence: 63%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Brain signatures for neuropsychological and everyday memory achieve high replicability and explanatory power in two data cohorts

Fletcher

Farias

DeCarli

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…In addition to issues arising from traditional approaches to segmenting brain IDPs, recent commentaries have pointed out inherent limitations from modeling outcomes associated with single brain measures from pre-selected lists 20 . The essence of this critique is that portions of multiple IDPs may be associated with an outcome of interest (e.g., cognitive performance or genetic association), crossing traditional ROI boundaries to recruit subsets of multiple regions while not using the entirety of any region.…”

Section: Mainmentioning

confidence: 99%

“…The essence of this critique is that portions of multiple IDPs may be associated with an outcome of interest (e.g., cognitive performance or genetic association), crossing traditional ROI boundaries to recruit subsets of multiple regions while not using the entirety of any region. Modeling brain-outcome associations with single whole IDPs thus loses both anatomical specificity (by forcibly incorporating more of a single IDP than is really associated to outcome) and sensitivity (failing to incorporate portions of other IDPs that are also associated) 20 . This is true a fortiori in GWAS, due to the pleiotropy or distributed influence of a genetic locus (SNP) on multiple brain regions and across different imaging modalities 21 .…”

Section: Mainmentioning

confidence: 99%

New phenotype discovery method by unsupervised deep representation learning empowers genetic association studies of brain imaging

Patel

Xie

Yuan

et al. 2022

Preprint

View full text Add to dashboard Cite

Understanding the genetic architecture of brain structure is challenging, partly due to difficulties in designing robust, non-biased descriptors of brain morphology. Until recently, brain measures for genome-wide association studies (GWAS) consisted of traditionally expert-defined or software-derived image-derived phenotypes (IDPs) that are often based on theoretical preconceptions or computed from limited amounts of data. Here, we present an approach to derive brain imaging phenotypes using unsupervised deep representation learning. We train a 3-D convolutional autoencoder model with reconstruction loss on 6,130 UK Biobank (UKBB) participants' T1 or T2-FLAIR (T2) brain MRIs to create a 128-dimensional representation known as endophenotypes (ENDOs). GWAS of these ENDOs in held-out UKBB subjects (n = 22,962 discovery and n = 12,848/11,717 replication cohorts for T1/T2) identified 658 significant replicated variant-ENDO pairs involving 43 independent loci. Thirteen loci were not reported in earlier T1 and T2 IDP-based UK Biobank GWAS. We developed a perturbation-based decoder interpretation approach to show that these loci are associated with ENDOs mapped to multiple relevant brain regions. Our results established unsupervised deep learning can derive robust, unbiased, heritable, and interpretable endophenotypes from imaging data.

show abstract

Toward a statistical validation of brain signatures as robust measures of behavioral substrates

Fletcher

Farias

DeCarli

et al. 2023

Human Brain Mapping

View full text Add to dashboard Cite

The “brain signature of cognition” concept has garnered interest as a data‐driven, exploratory approach to better understand key brain regions involved in specific cognitive functions, with the potential to maximally characterize brain substrates of behavioral outcomes. Previously we presented a method for computing signatures of episodic memory. However, to be a robust brain measure, the signature approach requires a rigorous validation of model performance across a variety of cohorts. Here we report validation results and provide an example of extending it to a second behavioral domain. In each of two discovery data cohorts, we derived regional brain gray matter thickness associations for two domains: neuropsychological and everyday cognition memory. We computed regional association to outcome in 40 randomly selected discovery subsets of size 400 in each cohort. We generated spatial overlap frequency maps and defined high‐frequency regions as “consensus” signature masks. Using separate validation datasets, we evaluated replicability of cohort‐based consensus model fits and explanatory power by comparing signature model fits with each other and with competing theory‐based models. Spatial replications produced convergent consensus signature regions. Consensus signature model fits were highly correlated in 50 random subsets of each validation cohort, indicating high replicability. In comparisons over each full cohort, signature models outperformed other models. In this validation study, we produced signature models that replicated model fits to outcome and outperformed other commonly used measures. Signatures in two memory domains suggested strongly shared brain substrates. Robust brain signatures may therefore be achievable, yielding reliable and useful measures for modeling substrates of behavioral domains.

show abstract

A robust brain signature region approach for episodic memory performance in older adults

Abstract: This scientific commentary refers to ‘A robust brain signature region approach for episodic memory performance in older adults’ by Fletcher et al. (doi:10.1093/brain/awab007).

Cited by 4 publications

References 8 publications

Brain signatures for neuropsychological and everyday memory achieve high replicability and explanatory power in two data cohorts

Brain signatures for neuropsychological and everyday memory achieve high replicability and explanatory power in two data cohorts

New phenotype discovery method by unsupervised deep representation learning empowers genetic association studies of brain imaging

Toward a statistical validation of brain signatures as robust measures of behavioral substrates

Contact Info

Product

Resources

About