G H PATEL COLLEGE OF ENGINEERING AND
TECHNOLOGY
(A constitute college of Charutar Vidya Mandal
University)
Coginitive computation
Prepared By :PART -3 Guided By :
Nisarg Jignesh Panchal ( CP ) Prof. Khyati Mehta
12302040501035
QUICK SUMMARY OF THE PREVIOUS TOPICS
• Introduction to BCI
• Working Model of BCI
• Applications of BCI
• How BCIs Use AI & ML
• Neural Correlates of Consciousness (NCC)
• Advanced Concepts: B2B Communication, Neural Plasticity,
Imprinting & Encryption
Key Computational Frameworks in Brain-Computer
Interfaces
ARTIFICIAL INTELLIGENCE (AI)
1. Neural Signal Processing Pipeline
ARTIFICIAL INTELLIGENCE (AI)
2. Neural Encryption Algorithm
3. BCI Decision Tree for Intent Recognition
4. Thought-to-Text Translation Process
ARTIFICIAL INTELLIGENCE (AI)
Neural Signal Processing Pipeline
PSEUDO CODE
ARTIFICIAL
INTELLIGENCE
(AI)
EXPLAINATION
1. Acquire raw EEG brain signals
2. Apply bandpass filter to remove noise
3. Extract meaningful features from the signal
4. Classify the features to predict user intent
5. Execute action based on the prediction
Neural Encryption Algorithm
ARTIFICIAL
INTELLIGENCE
(AI)
EXPLAINATION
1. Acquire EEG brain data from the user
2. Preprocess the signal (e.g., filtering, normalization) before
encryption
3. Generate a unique key using neural features or biometric
patterns
4. Encrypt the preprocessed signal using the generated key
5. Transmit the encrypted signal securely to the receiver/device
BCI Decision Tree for Intent Recognition
A decision tree is a supervised machine learning algorithm used
for classification and regression tasks. It's a flowchart-like
structure where:
• Each node represents a feature (or attribute),
• Each branch represents a decision rule,
• Each leaf represents an outcome (or label).
In BCI, a decision tree might be used to classify EEG
(electroencephalogram) signals based on extracted features.
BCI Decision Tree for Intent Recognition
Thought-to-Text Translation Process
• Thoughts generate unique brainwave patterns, especially when imagining speech or
specific words.
• EEG or neural interfaces record these brain signals non-invasively (or invasively in
some cases).
• Preprocessing and feature extraction clean the signals and highlight useful
information like frequency or spatial patterns.
• AI models (like neural networks) are trained to recognize patterns linked to specific
words, phonemes, or characters.
• The model predicts text by matching new brainwave inputs to learned patterns and
outputs the corresponding word.
• This enables hands-free communication, especially for people with disabilities, using
just their thoughts..
THANK YOU