BraiNP* - AI Based Predictive Models
# |
Diagnostic Models |
Specification |
Train Accuracy (%) |
|
| 1 |
Obsessive Compulsive Disorder <> Healthy Control
|
125 Hz. / 19 Ch. / 2m 58s |
98.5 |
Select |
| 2 |
Bipolar Disorder <> Health Control
|
125 Hz. / 19 Ch. / 8m 32s |
98.5 |
Select |
# |
Prognostic Models |
Specification |
Train Accuracy (%) |
|
| 1 |
Treatment Responder <> Non-Responder (MDD)
|
125 Hz. / 19 Ch. / 2m 58s |
98.2 |
Select |
*Many components of the healthcare delivery system include complicated and extensive decision-making processes that necessitate numerous
considerations before deciding on a course of action in patient care. Today, the opinions of each healthcare stakeholder are increasingly being considered,
which frequently leads to uncertainty and conflict in decision-making. BraiNP is a decision-making infrastructure settled on the know-how of Uskudar University
and NP Istanbul Brain Hospital. BraiNP is a predictive architecture with diagnostic and prognostic model components.
Diagnostic models use a deep dearning-based frameworks for feature extraction and classification.
While the inputs are the pre-processed neuroimaging raw data the output of the predictive model is the classification accuracy for the classes.
Prognostic models, employ deep learning-based structure to set a model aims to predict the risk of a particular future outcome for a specific subject for a specific
treatment. We believe that the interface with its structure and vision will be presenting promising results and great potential for precision medicine applications in
clinical research and practice.
