Try the new preview of the ThinkGear SDK for .NET - Mental Effort and Familiarity Beta.
What do Mental Effort and Familiarity tell about my mind?
Mental Effort measures the mental workload while performing a task. The harder you (your brain) work on a task, the higher the value. It could tell you many things such as – is your brain having a difficult time with the task?
Familiarity measures the process of learning while performing a task. For some cases, it reflects how well you are doing with the task. By observing trends, you can better understand your own learning process.
Both of them could be applied on tasks that are physical (e.g. drawing) or mental (e.g. recitation) in nature.
I would like to try out how they work first.
Our Demo Apps are located in the Demo folder of this SDK. Explore the executable demo applications and then try integrating the .NET code into your project.
Can you provide more example tasks that we can try on with Mental Effort and Familiarity?
We have compiled some of the tasks below:
Memorization: Poker card memorization, Random number memorization
Movements: fighting games, American Sign Language practicing, Mirror tracing, coloring pages, We have also conducted many experiments on those tasks – and Mental Effort and Familiarity provide thorough insight.
(See \docs\Algorithm Docs\Mental Effort and Familiarity\Experiment Examples and Test Results\).
Do note that the algorithms might or might not be applicable to any tasks you come up with outside our example tasks.
What is the key to reading the data? Are there any ways to interpret them?
Mental Effort and Familiarity have been released in beta. During this time, we are continuing to take in feedback and make refinements to improve the reliability of these algorithms. Please note that these algorithm are based on determining a status from your brain - refining data through comparisons. Within the beta, there is a 1-minute time boundary where the first minute of data should be used only for larger-scale trend observation and not be included within your dataset. If there are any future changes to this condition, it will be detailed here.
Please read the document “Reading Mental Effort and Familiarity.pdf” in the current folder. Additionally, your understanding of the algorithm can be reinforced by your own experience using the algorithms (e.g. while trying out our demo apps).
Additionally, you can conduct inter-trial (e.g. between “this time” and “that other time”) or intra-trial (e.g. between “this moment” and “that moment” in the same trial) comparison with the data obtained. Both can be done by calculating the percentage change of every obtained index from the baseline value (the first index from the trial - representing the relaxing session of 60 seconds):
Percentage change at the ith index = (the ith - baseline value) / baseline value
For the case of inter-trial comparison, you could compare a series of percentages changes from another such series, while both series are derived from a common baseline. The common baseline concept is further mentioned in the upcoming content.
The Mental Effort Analyzer is a secondary algorithm that helps inter-trial comparison. It helps summarize every trial into 2 figures (Total Effort, Effort Variability) in addition to the series of indexes (or percentage changes) you already have.
I know what to expect in the data now, but why are the values in arbitrary units?
The values are in arbitrary unit. You can see the fluctuation with real numbers as well. For example, for a series of Mental Effort Indexes collected while doing a task, even without a Y-axis, you see the fluctuation of effort over time.
What is it about “Idle Profiles” and/or “common baselines” that are mentioned in the materials?
Every Idle Profile contains some EEG data corresponding to a person that have been tested by the Mental Effort and/or Familiarity algorithm(s). The EEG data stored records the relaxed state (eyes open, doing nothing) of the person’s brain. The Idle Profile serves as a reference point for Mental Effort and Familiarity normalization.
Our implementation of the Idle Profile features in the SDK allows you to re-use the profile again and again – thus relieving the same person of repeated sessions of 1-minute idleness.
The Idle Profile also provides a clean start for both algorithms, and thus has to been plugged in to the start of a session whenever you are needing a fresh start in your app (e.g. Your app refreshing the task without closing and re-opening).
If you are digging into our materials, you may come across the term “common baseline”. The “common baseline” basically equates to the “Idle Profile”, where “common baseline” is a more scientific term.
What is the concept behind Familiarity?
Familiarity means learning (depending hugely on the task). When we define learning, we have to define the ability that is being learnt. For example, when a kid works on his penmanship homework, the ability is his/her motor ability. When someone is memorizing the set of road signs, the ability is his/her memorization of signs.
Another point is that learning can be stunned – we are constantly challenged by various aspects of a task, and each challenge affects our Familiarity values.
That brings us to: the Interpretation of Familiarity indexes is not as intuitive as Mental Effort indexes. Whereas Mental Effort can be read as a straightforward measurement of a subject's mental load at any given time, Familiarity must always take the task and the user's engagement with the task into account as described above.
Are there research papers to support your algorithms?
They’re at \docs\Algorithm Docs\Mental Effort and Familiarity\Research Papers\. Both papers involve studies that measures frontal EEG activities by a single-channel mobile EEG system. The paper titled “Evaluation of Mental Workload in Visual-Motor Task: Spectral Analysis of Single-Channel Frontal EEG” showed correlation between overall perceived difficulty of a visual-motor task and significant modulation on the frontal EEG spectra. Another paper titled “During Motor Skill Acquisition: Task Familiarity Monitoring Using Single-Channel EEG” showed correlation between overall familiarity level of the task and the frontal EEG activities in delta band of the whole trial and gamma band at the beginning of each trial.
In addition, we have various experiments done (see \docs\Algorithm Docs\Mental Effort and Familiarity\Experiment Examples and Test Results\). The document titled “Applying Mental Effort and Familiarity Algorithms in Maze Games” involves various Mental Effort and Familiarity results measured when the subject explore Mazes of various difficulties. Another document titled “Subjective Testing of Mental Effort and Familiarity Algorithms with Video Games” documented the tests on the usage model of Mental Effort and Familiarity algorithms on different video games such as “Coloring Pages” and “Street Fighter”. The experiments conducted confirm the prowess of the algorithms and help us understand more about the behavior of the data.
I’m a developer. Where do I start?
You should refer to the development guide that comes with your .NET / Android / iOS SDK package (e.g. ThinkGear SDK for .NET - Development Guide and API Reference.pdf). API references related to these algorithms are either under the sections API Reference/Connector class and API Reference/Algorithms.MentalEffort class in the development guide (.NET), or in the “reference” folder (Android). Technical descriptions and a step-by- step implementation guide can be found in the sections ThinkGear Data Types/EEG/MENTAL EFFORT and ThinkGear Data Types/EEG/FAMILIARITY.
We also have sample C# / Java / iOS projects demonstrating a basic implementation of Mental Effort and Familiarity (with data collection and basic analysis). It could be found in \Sample Projects and Demos\HelloMEandF\.
The Mac platform is forthcoming.
I want to know more. Who should I contact?
You can reach us at [email protected]
For your reference, you may check out our Developer Site at: http://developer.neurosky.com/