- AI in Cognitive Neuroscience and Computational Brain Imaging Research
- Neuro Informatics and Computational Brain Imaging
- Key AI techniques used in Computational Brain Research
- Cognitive Skills of Human Brain
- Significance of Working Memory
- Auditory Working Memory
- Visual Working Memory
- ADHD and ASD
- Role of Artificial Intelligence and Machine Learning
- Machine Learning Use cases on Working Memory
- Deep Learning Use cases on Working Memory
It is well known in today’s world that Machine learning has penetrated across all medical segments. This article explores how Artificial Intelligence, Machine learning (ML) and Deep Learning (DL) techniques impacts Cognitive Neuroscience and Computational Brain Imaging Research.
Neuro Informatics and Computational Brain Imaging
Cognitive Neuroscience deals with measurement of brain activity during cognitive tasks. Malfunction during basic cognitive process is due to few irreversible abnormalities in brain structure and improper neural connectivity and it leads to mental disorders. Neuro Informatics and Computational Brain imaging research use AI Predictions techniques to acquire human brain data and build mathematical models to simulate and investigate brain function abnormalities in early stages.
Key AI techniques used in Computational Brain Research
- Reinforcement Learning
- Natural Language Processing
- Speech and Language Processing
- Computer Vision
- Deep Learning Neural Networks
Cognitive Skills of Human Brain
Key Cognitive capacities of a normal human brain are
- Reading, Listening, Thinking, Decision making
- Learning, Logical reasoning, Complex Attention
- Executive Functioning – Metacognition skill
- Working Memory (WM) – The brain has the capacity to retain information for a short period of time using Visual and Auditory capability. The Cerebral Cortex of Human brain is highlighted as the Neural basis of Working Memory.
- Emotion Skills – Inhibitory Control and Self-Monitoring skills
- Self-Organizing Skills – Planning\ Priority\ goal directed behavior skills
- Cognitive Flexibility Skills – Problem Solving and multitasking
Significance of Working Memory
In rest of sections, we will focus on ‘Working memory executive skill’ as a significant AI use case.
Broadly, human Working Memory (WM) is classified as Visual and Auditory Working Memory.
Auditory Working Memory
Remembering the spoken words, keeping sounds in mind for short period of time when that sound is no longer present, remembering planned oral response, recognizing and responding the given oral instructions are some of the basic human ‘Auditory Working Memory’ or Verbal WM skills.
Visual Working Memory
Remembering human faces, remembering addresses or locations till reaching it, coping text from board to notebook in a classroom, remembering mathematical symbols during problem solving are some of the basic human ‘Visual Working Memory’ skills. When a visual stimulus is given the occipitotemporal cortex, Dorsolateral prefrontal cortex and Intraparietal sulcus plays important role in Visual WM encoding, maintenance, and retrieval.
Example Use case – Lack of Working Memory skill
ADHD and ASD
Attention Deficit Hyperactive Syndrome (ADHD) and Autism Spectrum Disorders (ASD) are mental disorders occurring in children which results in serious academic performance and behavior impact
According to Gathercole, et al., 2004 working memory is a key indicator of academic performance of school children in age group 6 to 14. Various Academic activities demand short time remembrance and recall such as reading, comprehension and problem solving. An academic classroom scenario has WM based tasks such as copying content from board, listening to step-by–step teacher instructions, lecture note making, writing big paragraphs, comprehending big paragraphs and answering questions based on it, remembering friends’ names, matching friends names with faces and mandatory requirement to remember class homework.
Brain’s Fusiform region has direct relationship with behavior impact. In ADHD patients we have visual and auditory working memory issues along with executive functioning impact. This is due to both structural and functional connectivity abnormalities in particular brain regions
Autism Spectrum Disorder (ASD) can be broadly divided into Low Functioning Autism (LFA) and High Functioning Autism (HFA). Patients identified with LFA has IQ lesser (<85) than HFA type. ASD patients are impacted in Executive function and social behavior.
Role of Artificial Intelligence and Machine Learning
AI based classification and diagnosis methods are used in autism-ADHD comorbidity detection, diagnosis, and classification of these diseases from the Brain MRI, EEG images.
Machine Learning Use cases on Working Memory
- Brain Functional connectivity Analysis of WM
- Brain region activation studies
- Brain Structural abnormal connectivity Analysis
- Brain Voxel based analysis and classification for given WM load
- Relationship analysis of WM cognitive processes.
- Analysis of relationship between WM and Behavior difficulty pattern analysis
- fMRI based WM encoding and decoding studies
- Study on Effect of domain -expertise on WM processes
- Study of Multivariate Pattern Analysis (MVPA) for Auditory stimuli brain activations and activity patterns
- Automatic feature selection from brain region of Interest for given WM task
Deep Learning Use cases on Working Memory
- Convolutional Neural Network (CNN) based Working Memory deep learning model Building
- Recurrent neural networks (RNN) based Brain Memory Architecture and mind model building
- Recurrent LSTM based Visual reasoning memory model
- Sensory and Semantic Decoding Memory model
- RNN based Working memory, forget and Long-term memory model
The core objective of Artificial Intelligence is to simulate human brain in terms of all it cognitive Processes and disorders (Memory, Visual Processing, executive functioning, logic and reasoning.) This article explained the key cognitive function ‘working memory’, impact of its deficit, AI/ML based use cases and solution directions.
 Fayyaz Ahmad1, Iftikhar Ahmad, Waqar Mahmood Dar, “Identification and classification of voxels of human brain for rewardless-related decision making using ANN technique” Natural Computing Applications Forum 2016, DOI 10.1007/s00521-016-2413-6
 M. Chiara Passolunghia and Linda S. Siegel,” Working memory and access to numerical information in children with disability in mathematics”, Journal of Experimental Child Psychology 88 (2004) 348–367,doi:10.1016/j.jecp.2004.04.002
 Sam Goldstein Jack A. Naglieri,” Handbook of Executive
Functioning” ISBN 978-1-4614-8105-8 ISBN 978-1-4614-8106-5 (eBook) DOI 10.1007/978-1-4614-8106-5