Search

link to homepage

Institute of Neuroscience and Medicine
(leer)

Navigation and service


Clinical Translation

A key goal of our research is to provide tools and platforms advancing the accuracy of diagnostic tools and outcome predictions through the application machine-learning approaches to imaging data and other clinical information.

TranslationThe medial frontal pole, which has previously been show to subserve socio-affective processing, shows a selective structural affection in major depression whereas the lateral frontal pole, which is part of a cognitive network, is unimpaired (Bludau et al., Am J Psychiatry 2016)

The investigation of brain-behavior relationships in patients suffering from brain diseases contributes critically to our research program through a bidirectional translation between basic and clinical neuroscience.

On one hand, the methods, tools and models developed in the INM-7 together with the normative data obtained from large population-based cohorts provide the necessary backdrop for characterizing system-level pathophysiology of the brain. Only through a systematic knowledge of normal brain variability, pathological deviations thereof may be accurately identified and used to characterize individual disease processes. Leveraging this knowledge to the benefit of individual patients moreover requires advanced approaches for statistical learning and big data analysis. In this context, a long-term goal of our research is to provide tools and platforms advancing the accuracy of diagnostic tools and outcome predictions through the combination of knowledge on human brain organization and variability into platforms for , and platforms for application machine-learning approaches to imaging data and other clinical information.

On the other hand, brain disorders also provide very important examples of network perturbations, providing insights into the relationship between brain structure, function and connectivity given the influence of pathological processes. In addition, the assessment of patients suffering from, e.g., major depression, schizophrenia or Parkinson's disease allow to extent our investigations of inter-individual variability in socio-affective, cognitive and motor-related performance beyond the physiological range. Complementing the assessment of large cohorts, clinical groups may thus provide additional information on structure-functionships in the human brain, reflecting other loss (Parkinson, stroke) or gain (Schizophrenia) of function.

Connectivity for different diseasesAccuracy for individual computer-based diagnosis of Schizophrenia (green), Parkinson's disease (blue) and aging (yellow) based on functional connectivity within 12 meta-analytically defined brain networks (left). Log-likelihood ratios of classification performance indicating that brain networks show a clear differentiation in the amount of information relative to the pathophysiology of Schizophrenia and Parkinson's, respectively (right).


Servicemeu

Homepage