File Server (NCILNAS)

The NCIL file server is called NCILNAS (for network attached storage). This is where all data should be stored. You should upload your research data to NCILNAS as soon as you have finished collecting the data. This is important because the file server is the only machine in the lab that is reliably backed up. Most other computers are not backed up at all; EEG data collection computers should be backed up to local external hard drives, but this is not as reliable. Keep in mind that all hard drives fail eventually, usually without warning, and always when it's most disastrous.

NCIL File Storage Policy

The following things should always be stored on the server. This is critical for us to fulfil our ethical and legal commitments, as well as our commitment to open, reproducible science.

Any and all files relating to experiments done in the lab

A folder on NCILscience should be created for each ethics protocol, and under this, all material related to this protocol, possibly including separate sub-studies in separate folders. This includes:

REB (ethics) documents

Include folders for each submission to an REB as well as a folder for communications from the REB, including letters of approval and annual renewals.

Stimuli and stimulus programs for the experiment

These are often created on experiment computers but should be uploaded to NCILscience before you start running the experiment, for backup and archival purposes. Stimulus computers are not reliably backed up. Within your Project folder on NCILNAS be sure to generate sub-folders for Stimuli and Data (in addition to Documents --> REB as per above). Within the Stimuli sub-folder seperate folders can be generated ( e.g., "DirectRT" or PyCorder for programming and other subfolders for actual stimuli such as videos or audio files, lists of words, etc)

All data collected as part of the protocol/experiment

All raw data shouldbe identified by subject ID code. All data should be stored on the server under the Project folder for Data,Note that any personally-identifying information must be stored separately from core study data that will be analyzed. This means in a separate file, but still under the same study folder. While our protocols differ, the only identifying information stored should be that absolutely necessary for the conduct of the study.

Unless otherwise required, contact information and other identifying information should be destroyed once data collection from an individual is completed. The TCPS definitions are as follows, “Information is identifiable if it may reasonably be expected to identify an individual, when used alone or combined with other available information. Information is non-identifiable if it does not identify an individual, for all practical purposes, when used alone or combined with other available information. The term “personal information” generally denotes identifiable information about an individual.”

Remember that most experiments include both EEG (saved on the EEG computer) or MRI data (stored on the scanner), AND behavioural data/stimulus program log files. All data files need to be copied to the server. I am shocked how often people completely forget to copy over the stimulus program log files - which can be invaluable in analyzing the results of your study.

All code used for data analysis

At the very least, all files necessary to recreate an analysis should be stored on the server at the time that a document based on those analyses is shared. This includes drafts of manuscripts shared with other people, and most certainly drafts submitted for publication (including posters) and published papers.

Results of all data analysis

In most cases data should live on the server, and for analysis you should connect to the server so that all code and results are also saved there. If for some reason you have to copy data to your own computer (please see Data Storage & Protection), be sure to copy the results back to the server after every analysis session.

All research outputs

• Theses

• Abstracts submitted to conferences

• Conference presentations (slides or poster)

• Manuscripts submitted for publication (in editable format, like Word or LaTeX; If you wrote the manuscript in Google Docs, you should export a version in Word format for storage on the server.)

• Proofs and preprints of manuscripts

• Final PDF versions of published papers

Backup Policy

NCILscience is backed up redundantly. No other computers in the lab are. That means that all behacioural, EEG, and fNIRS data collected on lab computers should be tranferred to the server. it is your responsibility to ensure that all research-related files are copied to NCILscience. This is tru even if you're using a service such as GitHub to develop and synchronize code (e.g., stimulus programs) across computers.

Remember, all hard drives fail. And rarely with any advance warning, or at a specific age. I have had drive fail days, weeks, or months after installation, and others that lasted for many years. Backing up should be your habit and religion.