AIoT based Neural Decoding and Neurofeedback for a Cognitive Training Acceleration
Ce projet a été attribué.
Encadrants
- Van-Tam Nguyen
- Emails: van-tam.nguyen@telecom-paris.fr
- Bureaux: --
Nombre d'étudiant par instance du projet:
- Minimum: 3
- Maximum: 5
Nombre d'instances du projet :
1Sigles des UE couvertes et/ou Mots-clés :
--Image
Description du projet :
The main objective of this project is to design adequate tiny machine learning and pattern
recognition algorithms in order to extract the relevant features and characteristics from
EEG and ECG signals, possibly implement them in microcontroller and finally design a
neurofeedback system to accelerate brain training (attention or/and working memory
or/and emotional intelligence) by improving efficiency and reducing training time.
Attention is a very important factor in cognitive efficiency. It allows us to notice and select
a subset of information from all that is available so that we can process that information.
Our attention system is needed for almost everything we do - whether it's learning,
memorizing, perceiving, communicating, or solving problems, etc. it is also important for
the regulation of our own emotions. Attention guides the allocation of processing
resources. Efficient resource allocation means rapid availability of information for superior
mental processing. Moreover, attentional guidance is crucial in tasks that require the
coordination of several cognitive operations by providing the appropriate resources. All of
these functions in information processing present attention as a major source of cognitive
efficiency. Finally, sustained attention and rapid changes between various cognitive
operations are strongly correlated with intelligence.
Working memory, another very important component of the cognitive base, emerged due
to mental activities requiring the availability of several pieces of information in a limited
amount of time. Such activities link multiple pieces of information together in a complex
pattern. It is essential for the mental activities that are believed to be the basis of
intelligence.
Because attention and working memory show a substantial relationship with intelligence,
and thus a strong correlation with academic and professional success, improving
attention and working memory is particularly relevant. Moreover, neuroplasticity is a
remarkable feature of the brain. Indeed, neurons are able to adapt quickly to the demands
imposed on them. By developing new neural networks and strengthening important
connections, a cognitive training program can measurably and sustainably improve brain
activity. It can also trigger the birth of new neurons. This is why the last two decades have
seen an impressive effort towards the design and implementation of cognitive training
programs especially with new technologies to improve general cognitive ability and slow
its decline in the elderly.
Objectifs du projet :
This project will not be to design a new computerized cognitive training technique, but
rather to build on millennium and well-known training techniques, such as mindfulness for
attention training and the method Loci or memory palaces for memory training. The main
objective is to collect appropriate physiological signals including EEG and ECG, to design
adequate tiny machine learning and pattern recognition algorithms to extract relevant
features and finally to design a neurofeedback system to accelerate the training of
attention and working memory by improving efficiency and reducing training time.
After signal acquisition, the main work consists of three main steps: feature extraction,
classification, and design of the neurofeedback system. The collected signals will be
analyzed both in the time domain and in the frequency domain. A nonlinear analysis will
also be performed. Tiny machine learning and pattern recognition algorithms from a
wearable device design perspective under limited hardware capacity constraints will be
investigated for feature extraction and classification. Finally, to validate the algorithms
studied, an implementation in Microcontroller will be carried out