Brain age as a surrogate marker for cognitive performance in multiple sclerosis

Denissen, Stijn; Engemann, Denis-Alexander; Cock, Alexander De; Baijot, Johan; Laton, Jorne; Penner, Iris‐Katharina; Grothe, Matthias; Kirsch, Michael; D'hooghe, Marie B; D’haeseleer, Miguel; Dive, Dominique; Mey, Johan De; Sima, Diana; Nagels, Guy

doi:10.1111/ene.15473

Cited by 31 publications

(43 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The root mean square error (RMSE) is To make the results comparable across the different traits, the RMSE is then divided by the range of the respective variable to obtain the NRMSE The second criterion, reliability, concerns two aspects: i) that the model will not show too large errors for some subjects that could harm interpretation; and ii) that the model’s accuracy will be consistent across random variations of the training set —in this case by using different (random) partitions for the CV folds. As mentioned, this is important if we want to interpret prediction errors for example in clinical contexts, which assumes that the error size of a model in a specific subject reflects something biologically meaningful, e.g., to what extent this subject defies the stereotypical population (Greene et al, 2022) or whether a certain disease causes the brain to “look” older to a model than the actual age of the subject (Denissen et al, 2022). While average errors can be informative to understand the general model performance, maximum errors inform us about single cases where a model (that may typically perform well) fails.…”

Section: Methodsmentioning

confidence: 99%

“…Second, they should be reliable, in the sense that a predictive model should not produce excessively large errors, and the outcome should be robust to the choice of which subjects are included in the training set. The latter criterion is especially important if we want to be able to meaningfully interpret prediction errors, e.g., in assessing brain age (Cole & Franke, 2017; Denissen et al, 2022; Smith et al, 2019). Despite this crucial role in interpreting model errors, error variance is rarely taken into account in models predicting phenotypes from neuroimaging features in the literature.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Predicting individual traits from models of brain dynamics accurately and reliably using the Fisher kernel

Ahrends

Vidaurre

2023

Preprint

View full text Add to dashboard Cite

How specifically brain activity unfolds across time, namely the nature of brain dynamics, can sometimes be more predictive of behavioural and cognitive subject traits than both brain structure and summary measures of brain activity that average across time. Brain dynamics can be described by models of varying complexity, but what is the best way to use these models of brain dynamics for characterising subject differences and predicting individual traits is unclear. While most studies aiming at predicting subjects’ traits focus on having accurate predictions, for many practical applications, it is critical for the predictions not just to be accurate but also reliable. Kernel methods are a robust and computationally efficient way of expressing differences in brain dynamics between subjects for the sake of predicting individual traits, such as clinical or psychological variables. Using the Hidden Markov model (HMM) as a model of brain dynamics, we here propose the use of the Fisher kernel, a mathematically principled approach to predict phenotypes from subject-specific HMMs. The Fisher kernel is computed in a way that preserves the mathematical structure of the HMM. This results in benefits in terms of both accuracy and reliability when compared with kernels that ignore the structure of the underlying model of brain dynamics.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Predicting individual traits from models of brain dynamics accurately and reliably using the Fisher kernel

Ahrends

Vidaurre

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…In contrast to a knowledge-based representation, this latent space is a data-driven representation, which is typically not interpretable for humans. Although it is unclear whether these representations are a useful alternative to overcome the aforementioned paradox, we recently showed that brain age is related to disease burden of persons with MS in terms of information processing speed, independently of their chronological age (8). Analogously to Leonardsen et al 2022 (7), we will now use transfer…”

Section: Introductionmentioning

confidence: 91%

Transfer learning on structural brain age models to decode cognition in MS: a federated learning approach

Denissen

Grothe

Vaněčková

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Introduction: Classical deep learning research requires lots of centralised data. However, data sets are often stored at different clinical centers, and sharing sensitive patient data such as brain images is difficult. In this manuscript, we investigated the feasibility of federated learning, sending models to the data instead of the other way round, for research on brain magnetic resonant images of people with multiple sclerosis (MS). Methods: Using transfer learning on a previously published brain age model, we trained a model to decode performance on the symbol digit modalities test (SDMT) of patients with multiple sclerosis from structural T1 weighted MRI. Three international centers in Brussels, Greifswald and Prague participated in the project. In Brussels, one computer served as the server coordinating the FL project, while the other served as client for model training on local data (n=97). The other two clients were Greifswald (n=104) and Prague (n=100). Each FL round, the server sent a global model to the clients, where its fully connected layer was updated on the local data. After collecting the local models, the server applied a weighted average of two randomly picked clients, yielding a new global model. Results: After 22 federated learning rounds, the average validation loss across clients reached a minimum. The model appeared to have learned to assign SDMT values close to the mean with a mean absolute error of 9.04, 10.59 and 10.71 points between true and predicted SDMT on the test data sets of Brussels, Greifswald and Prague respectively. The overall test MAE across all clients was 10.13 points. Conclusion: Federated learning is feasible for machine learning research on brain MRI of persons with MS, setting the stage for larger transfer learning studies to investigate the utility of brain age latent representations in cogni- tive decoding tasks.

show abstract

“…By using large datasets of brain images of healthy individuals, a machine learning model can be trained to estimate the age of a given brain [ 86 ]. The model can provide the best guess of the age of that person’s brain, that is, the ‘brain age’, which can look older or younger than the actual, chronological age of that person.…”

Section: Clinical Significance Of Demyelinationmentioning

confidence: 99%

Role of Demyelination in the Persistence of Neurological and Mental Impairments after COVID-19

Khodanovich

Parshukova

Naumova

2022

IJMS

View full text Add to dashboard Cite

Long-term neurological and mental complications of COVID-19, the so-called post-COVID syndrome or long COVID, affect the quality of life. The most persistent manifestations of long COVID include fatigue, anosmia/hyposmia, insomnia, depression/anxiety, and memory/attention deficits. The physiological basis of neurological and psychiatric disorders is still poorly understood. This review summarizes the current knowledge of neurological sequelae in post-COVID patients and discusses brain demyelination as a possible mechanism of these complications with a focus on neuroimaging findings. Numerous reviews, experimental and theoretical studies consider brain demyelination as one of the mechanisms of the central neural system impairment. Several factors might cause demyelination, such as inflammation, direct effect of the virus on oligodendrocytes, and cerebrovascular disorders, inducing myelin damage. There is a contradiction between the solid fundamental basis underlying demyelination as the mechanism of the neurological injuries and relatively little published clinical evidence related to demyelination in COVID-19 patients. The reason for this probably lies in the fact that most clinical studies used conventional MRI techniques, which can detect only large, clearly visible demyelinating lesions. A very limited number of studies use specific methods for myelin quantification detected changes in the white matter tracts 3 and 10 months after the acute phase of COVID-19. Future research applying quantitative MRI assessment of myelin in combination with neurological and psychological studies will help in understanding the mechanisms of post-COVID complications associated with demyelination.

show abstract

Brain age as a surrogate marker for cognitive performance in multiple sclerosis

Cited by 31 publications

References 47 publications

Predicting individual traits from models of brain dynamics accurately and reliably using the Fisher kernel

Predicting individual traits from models of brain dynamics accurately and reliably using the Fisher kernel

Transfer learning on structural brain age models to decode cognition in MS: a federated learning approach

Role of Demyelination in the Persistence of Neurological and Mental Impairments after COVID-19

Contact Info

Product

Resources

About