Version 1.1

Version 1.1.9, 06 Apr 2009

Relaxed record validity check so that JMWork would accept XWork-processed files affected by a benign bug that introduced non-zero bytes in bytes 2-7 of the 8-byte header of certain XWork-added records.


Version 1.1.8, 17 Mar 2009

    • No longer automatically turns on the display of a sorted spike train channel if data file contains data for such a channel -- because this overrides the user's preference.

    • Viewport start time can be specified as an absolute time or as the time of occurrence of a digital event pulse.

    • Highlight region end time t1 can now be specified -- like the start time t0 -- as the time of occurrence of a digital event pulse.


Version 1.1.7, 16 Jan 2009

    • New feature: If you check the Enable auto-discard feature box in the Saccade Cuts panel, JMWork will automatically mark the data file as "discarded" if a saccade is located within the highlight region. A saccade is defined either by cuts in the eye-velocity traces (XWork-style), or by mark1/2 pairs (MWork style), depending on the user's preference. Basically, this saves a keystroke if your goal is to reject trials in which the animal saccades during a key segment in the trial.

    • Added another hot key: U removes the discard mark (same effect as Shift-Q).

    • When you filter the file list by filename, the set of unique trial names in the Filter By Trial combo box is updated to include only trials from the filtered file set.

    • Minor appearance changes: a splashscreen at startup, a different application icon, and a different cursor for "adjust" mode.


Version 1.1.6, 12 Dec 2008

    • No longer "wraps back" to the beginning when you step past the end of the file list.

    • No longer removes all marks/cuts when you "discard" a file.

    • Added hot keys: Q to mark a file as discarded without stepping forward, Shift-Q to remove the discard mark (IAW the current discard-mark style!), and spacebar to step forward.


Version 1.1.5, 14 Nov 2008

When reviewing results in the Analysis Tasks dialog, one can select a trace on the task results canvas by simply clicking on or near it. This should make it easier to pick out a trace that you need to exclude from the saved results.


Version 1.1.4, 13 Oct 2008

    • Added File menu command Refresh Data Directory. This reloads the current data directory's file list from scratch, ignoring any cached content for that directory. Before this change, JMWork would always use the cached information to populate a data directory, which rendered it useless for examining data files collected while an experiment is in progress.

    • Added a second means of deleting a single saccade cut, mark segment, mark1/2 pair, or isolated mark: Right-click on the relevant row in the Cuts and Marks table, with the Shift key depressed.


Version 1.1.3, 01 Oct 2008

Changed appearance of cursor in spike-edit mode. It now looks like a lightning-bolt, with accents to indicate what operation will occur given the current cursor location. If the cursor is inside a sorted-spike channel band and: (a) is within +/-3 pixels of an existing spike edit, an "undo spike edit" accent appears (a curved green arrow); (b) is within +/-3 pixels of a spike in the unedited spike train, a "remove spike" accent appears (a red 'X'); (c) otherwise, an "add spike" accent appears (green "+"). If the cursor is inside the band in which the current spike sorter results are drawn, the accent is a magnifying glass -- since clicking near a spike here will highlight that spike in the spike sorter canvas. Otherwise, the lightning-bolt cursor takes on a "disabled" appearance -- more subdued color and no accent.


Version 1.1.2, 17 Sep 2008

    • Increased tolerance of cursor that edits individual spikes in spike-edit mode. Now, if an existing spike edit is within +/-3 pixels of the cursor position, it is removed. Otherwise, if there's a real spike within +/-3 pixels of the cursor, that spike is removed. If neither is the case, only then is a spike added at the cursor position.

    • Added a parameter for the spike-sort algorithm: hysteresis, the number of consecutive samples that must lie below a negative trigger level or above a positive trigger level to "detect a spike". In previous versions, this parameter was hard-coded at 4. Now the user can set it to any value in [1..5]. This significantly improved the spike-sort result for some data Steve was working on.


Version 1.1.1, 27 Aug 2008

    • Bug fix: When running JMWork remotely under Linux, it would take several seconds to update the trace display and spike sorter canvases. Apparently, this is because Java2D has to software-render graphics that involve translucency, which involves a whole lot of back-and-forth blitting over the network when running remotely. To circumvent the issue, JMWork now uses only opaque colors in the canvases. You may notice some slight color changes, but functionality is unaffected. And performance is now MUCH better when running remotely.

    • Revised, 08 Sep 2008: Modified analysis task-related code to account for a minor extension in the allowed character set for DataNav-compatible dataset IDs.


Version 1.1.0, 19 Aug 2008

    • [Leopard MacOSX-specific] Fixed layout problem which caused the canvas control panel to disappear when the frame window was resized.

    • Added the File|Analyze... command, which raises a modal dialog by which the user can perform data-directory wide analysis tasks. For now, only one type of analysis task is supported: trial data file collation and/or averaging. A user can define any number of these collation tasks; task definitions are saved in JSON format in the user's private JMWork directory. If users request additional kinds of tasks, the framework is now in place to develop such additional tasks relatively quickly.