Activities

No upcoming events

Intelligent Systems Lab Project: Brain @ Work: BRAWO

Participants

• Akshat Tandon
• Heiko Krause
• Stefan Weber
• Umakshi Sajnani
• Viswa Subramanian Sekar

Supervisors

• Dipl. Inform. Andrea Finke

Motivation

• A technology becomes successful only if it reaches the common people and makes the living environment a better place to live in with the efficient use of technology.
• The axiom of "Ambient Environment" is that it blends technology with the living atmosphere, and doesn't give the "techie" feel, as well as acting in an intelligent way, making the user acclimatized to the place.
• Brain Computer Interfaces are becoming more and more popular, because of its highly interdisciplinary research and the mind-blowing data that you can read from brain which can be used for various purposes in research.
• This project "Brain @ Work: Brawo" is an Intelligent human monitoring and feedback system.

Problem Statement

• Workplace stress has been reported to be a negative factor in human health and well-being. While several reasons can be attributed to the generation of stress, such as bad, improper working conditions or social stress.
• Our application tries to tackle two specific reasons for stress. Stress induced by working, when a pause would be appropriate and not being able to relax after working.
• In our application we are taking this one step further to not only track the mental activity but use a technique called Neurofeedback training to get the attention on spot again.

Results of the First and Second Term (Winter Term 2014/15 and Summer Term 2015)

Application Scenario

Our application collects EEG data and eyeblinks using the Mindwave Neurosky headset, ECG data using Bitalino and imports it into a common framework called UBiCI.

Basic Scenario

This framework analyzes the collected data and sends the results via RSB to a number of different modules:

• Neurofeedback Game
• Ambient Intelligent Room
• Smart Hues lamp
• Info Plant
• Music Player

Mind State Classification

Hardware Description

MindWave Neuro Sky

EEG Sensor

Certain electrical patterns of brain activity on the scalp are detected.

• Brain Mapping - Analysis of the digitized EEG.
• Ability to view dynamic changes in the brain during cognitive processing tasks.
• Spontaneous EEG - judges level of consciousness.
• Brain Waves characterized by frequency bands:
• Delta : 1-4 Hz - Deep Sleep.
• Theta : 5-7 Hz - Meditation.
• Alpha : 8-12 Hz - Relaxation.
• Beta : 13-30 Hz - Concentrated.
• Gamma : >30 Hz - Memory Processes.

Bitalino − ECG Sensor

ECG Sensor

• The bitalino board is able to work with several bio signal sensors. In this case we use the ECG (Electrocardiography) to get the sinus rythm of the heartbeat.
• Out of several features that could be calculated with that, the pNN50 and RMSSD have been chosen as significant representation for the heart rate variability (HRV).
• The parasympathetic nervous system controls the heart rate while being in rest and makes the heart beat less exactly rhythmic, as when the sympathetic nervous system takes over, which is associated with stressful situation.
• Thus the HRV can be used to pinpoint if a person is in a stressed or relaxed mood.
• The RMSSD is the square root of the mean squared difference between adjacent peak-to-peak intervals.

• The pNN50 is the percentage of adjacent peak-to-peak times, which differ more than 50 ms from each other.
• Both these values indicate rising stress as they decline.

Use case Scenario

We consider 2 use cases.

1. The 1st considers a working user, trying to avoid phases of non-productivity.
• The eyeblink rate as well as an activity index computed from the EEG data is then taken as an indicator of the tiredness of the user.
• Once it crosses a specific threshold, the music player notifies the user that he should take a break.
• An acoustic signal was chosen due to its non-intrusiveness. The user is now informed to take a break as soon as possible.

Mapping of Mind State for The NeuroFeedback Game
• To get the attention up again, the user can choose to play the Neurofeedback game. This uses a simple way of displaying the user his real time mental activity.
• He can use this feedback loop to train his ability to change mental states on a long timescale.
• As a short time effect, playing the game makes the user more focused or relaxed according to his needs.
2. The home automation is a silent way of indicating the mental state to the user. This is a constant but subtle feedback bringing our application in the realm of ambient intelligence.

Mapping of Mind State for The Ambient Environment
• The above figure shows the respective mind states mapped to the Philip hues bulb and the Info plant.
• The Hues Light as well as the Info Plant change color according to the user's concentration level. When the user reaches hyperactive state, the music player is turned on as an indication to relax.

The Ambient Environment: Various Features

Video

The following video describes all the features:

Evaluation

To test the Neurofeedback game, i.e whether it brings the user to the needed mindstate or not, 3 participants played the game for a span of 4 weeks with 3 trainings in a week. The game constituted of 2 modes, where the person has to be concentrated to finish the game or relaxed to finish the game. Results: the data we took is insufficient

• Didn't show the long term learning curve we hoped to get.
• But, everyone has a little higher result from first to last training.

Conclusion

• Mind state classification of human's through EEG supported by ECG established successfully.
• Controlling of hardware through the classified states.
• Establishing an Ambient environment.
• Future possible extension − real time environment testing.
• Neurofeedback training for more number of people and evaluation of the results

© 2014 Bielefeld University