Abstract The number of people living with Alzheimer’s disease (AD) is expected to increase from 46.8 million worldwide (5.2 million in the USA) at 2014 to 140 million worldwide (13.8 million in the USA) by 2050. The AD is caused by decline in cognitive function; hence, this frightening rate demonstrates the importance of assessing cognitive learning rate (the speed of cognitive performance improvement; change of performance score divided by practice time) as a diagnostic tool to prevent serious cognitive deterioration. Several subjective evaluation measures have been offered, but unrelated factors to cognitive state such as age, education, tester bias, and patient discomfort, may have an impact on individual ratings. Furthermore, multiple electroencephalogram (EEG) studies have shown that dementia patients have different brain functioning parameters than healthy people. However, inconsistent findings have been reported, and no universally accepted screening technique for early diagnosis of cognitive impairment exists. The goals of this study are to develop models that use high-density EEG recordings from 120 areas of the brain to objectively evaluate cognitive learning rate and cognitive performance status in healthy participants. The hypothesis is that functional brain network of individual areas of the brain changes during cognitive learning (learning how to perform cognitive tasks properly). To validate this hypothesis, the EEG signals of thirty mentally healthy participants (age: 36±12.7; 10 females and 20 males) performing three cognitive tasks will be used. These tasks have been created and implemented in a robot simulator framework to help surgical trainees enhance their cognitive skills. Perception, understanding, thinking, decision making, and task/time management are required for these tasks, as well as spatial cognitive capacity, which allows for a quick, correct knowledge of the position, orientation, size, and form of the object on which surgery is being performed. At the end of each task, the robot simulator delivers a performance score. Participants are told to repeat these tasks until they get a passing score. Then, computational network neuroscience algorithms will be developed to extract important features characterizing the whole-brain information processing efficiency. Multivariate analysis will be used to determine the correlations between the participants' cognitive learning rate and retrieved features. Using EEG signals, a deep neural network model will be developed to classify performance of completing cognitive tasks into three categories: failed (score <71), passed (score:71-80), and excellent (score: 81-100). Although the models in this work were constructed using EEG data from healthy people, they may be tweaked to fit data from AD patients completing more common cognitive activities. Once validated in the area of AD, the developed models could be used to monitor cognitive learning rate and cognitive performance to dis...