February 23, 2006
frequently asked questions FAQ
download PRsetup1212.exe to install latest public release (release February 23, 2006)
download PRsetup1055.exe to install stable beta version (release July 10, 2003)
The use of PolyRex is free, but it is not freeware. Any distribution and use requires the authorization from our group, which is typically accomplished via email.
It is required that any publication including data processed or manipulated by PolyRex will properly cite the use of this software in the Methods section as:
or in a similar standard format for web references as stated in most journal guidelines. Additional acknowledgments regarding free access to PolyRex would be appreciated, at the author's discretion. Please make sure that the head/director of your department/research group, and any other person who might publish findings stemming from the use of PolyRex, is aware of these conditions.
The download files are either an installation executable (since version 18.104.22.168) or a self-extractring zip archive. A password is needed to extract the included file(s). The current password is the same for all installations files/archives and will be forwarded on request by email: "email@example.com" (please state your name, institution, email address, and intended research purpose).
After installation or extraction to any directory, start the program PolyRex.exe. If missing, a new PolyRex.ini file is automatically generated. Point to a "Source" directory that contains the BDF file(s) (only files with a *.bdf extension are listed), select at least one source file, and adjust conversion settings. Some functions are also or only accessible via a context-dependent pop-up menu (use right-mouse click). Select destination format (CNT or BDF) and convert file(s).
To test PolyRex, you can download a zip archive with test data in BDF file format from BioSemi's web site (BDFtestfiles.zip 1.92 MB contains two short dummy data files) and from our NYSPI Psychophysiology web site (BDFerpdata.zip 2.24 MB stores one EEG recording with real ERP data including stimulus and response events).
|Authorization and Disclaimer
The copyright-protected software is provided 'as is' (the most recent version is usually considered being in a beta developing stage).Its sole use is intended for non-profit scientific research purposes. Since its original 2003 release, PolyRex has been used in the Psychophysiology Laboratory of New York State Psychicatric Institute (http://psychophysiology.cmpc.columbia.edu), and is evidently by now also used in many other research institutions. The software is revised if bugs or problems are identified, or other modifications are needed. However, all responsibilities and consequences of using this software are completely with the user (you should know what you are doing). At present, the documentation is limited to the brief description below, a growing, however, still rudimentary FAQ page, and various interface hints (although the interface should be largely self-explanatory). Please report bugs or any other problems to: firstname.lastname@example.org. I will try to address any issues if my time allows this. Although there is no charge or any other usage fee, the software is not 'freeware' or 'shareware' and may not be distributed without explicit permission of the author. It is presumed that use of this software for any published research, or implementation of original ideas outlined below and the program's About box, will be acknowledged by appropriately citing this software (see Conditional Use above). By downloading and installing this software you are agreeing with these terms.
Software Background and Overview
PolyRex was originally developed for the sole purpose of down-converting the 24-bit BioSemi data format (BDF) used by BioSemi's ActiveTwo EEG/ERP recording system to the 16-bit continous data format (CNT) used by NeuroScan (version 3.x) with minimal loss in data resolution. Although a converter program had been distributed by BioSemi under subcontract, its conversion results did not fully meet our data collection and analysis needs, which originate from a thorough understanding and exploitation of NeuoScan's 3.x data format. It was easier for us to write our own software than to request and explain the needed modifications for an effective integration of ActiveTwo data recordings into our existing data processing and analysis procedures.
PolyRex was written in Pascal (Borland Delphi 7) under Windows OS, and has been tested under Windows XP and ME (Pentium III/IV, RAM 128-512 MB) using Scan 4.1, 4.2, 4.3 and Edit 3.2.43 (results appear to work also with older versions of Edit). Several basic ideas were implemented in PolyRex. First, for all practical EEG/ERP purposes, the storage of "reference-free" data (see BioSemi's documentation for using an active reference with a passive electrode) is no longer needed after data collection, and in fact meaningless until the data have been rereferenced. Thus, PolyRex analyzes the digital data range per recorded channel, after (optionally) applying any desired reference montage (e.g., Cz, linked mastoids, average, etc.), and removes the (meaningless) offset from each channel (these values can be stored in unused variables of the NeuroScan header, or exported to an ASCII file). This procedure effectively minimizes the loss of data resolution due to fitting a 24-bit integer range into a 16-bit integer range. Second, several BDF files can be converted as a group (e.g., different blocks of an experiment), for which digital ranges and their offset corrections are then calculated across files, thereby maintaining the same data resolution in these different files. Third, trigger events (the last BDF channel contains one byte for stimulus events, one byte for response events, and one byte for system recording events) may be decoded as used in the hardware setup. For instance, any subrange of the bit sequences may be used, and bytes may be selectively inverted. Fourth, to improve conversion results, unused and bad channels may be excluded or flagged, and offset shifts at recording interruptions can be eliminated. Fifth, new channels may be computed from the data of the recorded channels, and added to the conversion montage. For example, our hardware setup has additional face (monopolar) leads at the outer canthi of each eye and at supra- und infraorbital sites of the right eye, which are used to compute horizontal bipolar and vertical bipolar eye movements. Sixth, any occurences of data saturation (data overflow and underflow) are reported for both the original data recording (theoretically impossible for the 24-bit storage format, however, possible due to hardware/amplifier saturation), and for the converted data (possible depending on the selected options).
last updated: June 3, 2013