FREE ACCESS TO DATA CLEANING AND DATA MINING SYSTEM

To obtain access to the data cleaning and data mining system, please complete the Online Form to sign up as an alpha tester. Once you have submitted the form, we will create a user account for you so that you can access our system through the URDA software client available from the product download page. Once you have downloaded and installed the software, you can submit your data to our server for analysis.  This can be done using a free dropbox account which will enable you to work on your data as if you were doing the computation on your own computer (with some small file folder update time).

HOW TO USE THE SOFTWARE: 4 EASY STEPS

1. Data Upload– Copy your data into your dropbox directory (?upload your data)? and receive information describing your data.
2. Clean Your Data– Execute algorithms to clean the data you have uploaded to the server.  You may access (or download)? the cleaned data from the server via your dropbox directory along with figures and files describing the results of the cleaning process and how effectively your data were cleaned.  If you have no interest in mining your data, you can stop at this point or clean additional data. ??
3. Mine Your Data– If data mining is your purpose and if the data are determined to be satisfactory in steps (1) and (2), then the data mining algorithms can be used to construct a data-driven model of the brain activities contained in your data will be created.  According to the input you provide, comparisons will be made among experimental conditions to show differences in functional brain activity among conditions.
4. Download Results– Once data mining is completed, the result will be available to you for download.

PRIMARY BENEFITS OF REMOTE PROCESSING

There categorically 2 main benefits to remote data processing/analysis over traditional approaches. First, the URDA data cleanining and data mining system provides access to algorithms that you would not otherwise be able to access.  We have implemented multiple wavelet and ICA based algorithms that we ourselves have found useful in our own analysis.  Second, you do not have to invest in extra computer resources to process your data; some of these algorithms require a significant amount of processor time and memory as well as specialized software.  Thank-you for investing the time to try our URDA service and client software. Feel free to provide us feedback on your experience and how we might improve our system for future users.

CHARACTERISTICS THAT SET OUR DATA MINING METHODS APART

• Typically, greater than 15 (for 32 electrodes), 30 (for 64 electrodes) areas of the brain contributing activities to the EEG are identified
• Components are validated as “brain” based on the likelihood that these components relate to distinct, modular regions of the brain
• Components are not restricted to originate from the edge of the cortex under the scalp
• Analysis (from the user’s point of view) is largely automated
• No need for you to be an expert in the subtleties of the latest complex data mining processes
• Access to EEG analysis and study design expertise: we help you find what is contained in your data
• The ability to study complex EEG data related to complex human behavior
• Access to our sophisticated data cleaning and artifact marking algorithms so that you can keep as much data as possible (and not reject it because it contains an artefact that, with the right algorithm, can be removed from the data)

Researchers/Students:

EEG data cleaning is helpful in studies investigating brain function as a way to minimize between participant variance due to non-brain-related variance in the data. EEG data mining provides researchers a peek at what their data contain and can inform their next research steps. Mine your data before doing statistical tests or mine the data after completing a research project. The data mining results can provide information that will help complete your conceptual model of the process under study and guide you to your next research question.

Practitioners:

EEG data cleaning is useful for minimizing time scrutinizing datasets to pick non-artefact time segments of data for analysis.  Data cleaning is particularly useful for those cases where artefact exists at nearly all reasonably sized intervals in the dataset.  For example, artefact EEG is difficult to obtain from clients who incessantly ‘blink’, or clients who for some reason produce EEG with large heart-beat artefact.  Data cleaning can make some datasets analyzable that otherwise contain too much artefact to apply standard analysis methods.In addition to using QEEG methods for relating global EEG data to behavioral disturbances/disorders, EEG data collected while patients participate in cognitive testing paradigms can be related to linked to specific areas of the brain and to specific aspects of the testing paradigm. These EEG can be mined to provide insight into the cognitive strategies and underlying brain physiology employed while the EEG were collected.

EXAMPLE DATA MINING APPLICATIONS

EEG data mining is ideal for:

• Pilot studies for which the time location and relevant EEG channel or brain location is not precisely known
• Novel behavioral paradigms because it will inform the investigator what hypothesis tests to make using standard methods
• Examining EEG data collected under-task because it can successfully process data collected during complex human behaviors

Suitable paradigm tasks include:

• Stroop and Conflict Tasks: examining inhibition and executive function
• Change of Cognitive Strategy Tasks: identifying brain physiology related to strategy
• Treadmill Walking: examining cognitive aspects of compensation for PD
• Video Game Playing: examining cognition and brain physiology associated with game playing
• Visualization and Motor Imagery: developing imagined movement Brain-Computer Interface (BCI) applications. Identify which electrodes are most relevant to the brain activity of interested; identify outlier participants (left vs. right hemisphere dominance)

CONSULTATION: HOW WE CAN HELP

We highly recommend a we provide you a consult before collecting EEG data. Our experience has been that in most cases, small modifications to paradigm can greatly improve the quality of the data mining result. In addition, we will provide you with some tricks for improving data quality. If you are at all interested in using our consultation services for your investigation, contact us and we will make time for an online Skype meeting.