Such decoding procedures are used in neuroscience to detect brain activation patterns that characterize the underlying mental processes. To characterize the function of the recruited brain structures, it is now common to input these data to machine learning models, to infer mental processes from brain activity recordings 3. Standard statistical analysis is based on a subtractive logic: fMRI data are used to measure local differences in brain oxygen supply between two experimental conditions, a target condition that includes the cognitive concept of interest 1 and a control condition that is either a related task or a baseline 2. For instance, in task functional Magnetic Resonance Imaging (fMRI) studies, experimental protocols trigger cognitive processes in a set of participants, while the corresponding neural activity is recorded. It enables broad reverse inferences, that is, concluding on mental processes given the observed brain activity.Ĭognitive neuroscience probes the relationships between mental functions and brain systems by contrasting brain responses observed in conditions that involve these functions to control conditions.
#MENTAL PROCESSES ARE MANUAL#
This success demonstrates that image-based meta-analyses can be undertaken at scale and with minimal manual data curation. Overcoming the heterogeneity, imbalance and noise in the training data, we successfully decoded more than 50 classes of mental processes on a large test set.
We trained neural networks to predict these cognitive labels on tens of thousands of brain images. We labeled NeuroVault images with Cognitive Atlas concepts occurring in their associated metadata. For this, we leveraged two comprehensive resources: NeuroVault-an open repository of fMRI statistical maps with unconstrained annotations-and Cognitive Atlas-an ontology of cognition. Using all the functional Magnetic Resonance Imaging (fMRI) statistical maps of the largest data repository available, we trained machine-learning models that decode the cognitive concepts probed in unseen studies. These analyses typically face a difficult compromise between scope-from domain-specific to system-level analysis-and accuracy.
Associating brain systems with mental processes requires statistical analysis of brain activity across many cognitive processes.
0 kommentar(er)
