Version 2.0

Version 2.0.5, released 17 Jan 2007

    • Introduced additional special operations chooseFix1 and chooseFix2, which implement a straightforward forced-choice protocol: As soon as eye is within fixation window of the "right" target (Fix1 or Fix2), the "wrong" target is turned off, reward pulse #2 is delivered, and fixation checking is enabled for the rest of the segment. If the special segment ends and the "right" target was never fixated, the trial aborts. Implementation constraints similar to those of selByFix.

    • Bug fixed: In Test mode, got a debug assertion upon hitting the "-" key, whether on the numeric keypad or the regular minus key.


Version 2.0.4, released 13 Dec 2006

    • The Target Editor now supports four different modification modes, akin to what is already available on the Trial Editor. A pushbutton at the top-left corner of the form displays the current modification mode and switches among the modes. Keyboard shortcuts are also available (Alt+6 through Alt+9).

    • Added support in Continuous mode for moving the pattern of an "active target" at a constant velocity. Two additional parameters were added to the grid control in the Active Targets tab -- pattern speed in deg/sec and the direction of pattern motion in deg CCW from the positive x-axis. Global keyboard shortcuts allow the user to increment/decrement the pattern speed and direction for the Track target only. These features were primarily added to support the use of the Track target to assess the response properties of a neural unit.


Version 2.0.3, released 06 Dec 2006

    • To reduce the chances of accidentally selecting a NOFIX mode, reshuffled the list of trial sequencer modes so that all NOFIX modes are at the bottom of the dropdown list for the Seq mode combo box on the Protocol tab of the Trial mode control panel. Increased height of the dropdown list so that all choices are simultaneously visible.

    • Added keyboard shortcuts for starting/stopping (function key F7) or pausing/resuming (function key F8) the trial sequencer in Trial mode.

    • Added another modification mode to the Trial Editor: Modify Across Selected Trials in Set (keyboard shortcut Alt+5). In this mode, any change made to the current trial will be propagated to all compatible trials in the parent trial set that are currently selected in the document tree.

    • Fixed minor bug in RMVideo [If the previous trial paints targets on its very last frame, that frame appears at the beginning of the next trial b/c RMVideo fails to clear the backbuffer prior to painting the first frame of a new animation sequence.]

    • Updated the RMVideo Random-Dot Patch target so that it allows a noise offset range of [0..300%] instead of [0..100%] -- as was done for the XYScope target Noisy Dots (Speed) in v2.0.2.


Version 2.0.2, released 23 Jul 2006 (rev 24 Oct)

    • Minor change to the Noisy Dots (Speed) target: The allowed noise offset range, N, is now [0..300%] instead of [0..100%]. Of course, if you choose N>100, some of the individual dots will move "backwards" (ie, in the opposite direction of the mean target motion).

    • Revised 14 Sep: Corrected a mistake in the keyboard "tab order" of the controls on the Protocol tab of the Trial mode control panel.

    • Revised 27 Sep: Fixed bug in trial sequencer, involving preparation of trial codes when a target is accelerated. [Symptom: If target window velocity was V during segments N thru N+2, and target window acceleration was nonzero during segment N+1, target window velocity failed to return to V at the start of segment N+2].

    • Revised 24 Oct: Fixed pseudo-bug that impacted use of the plexmon Matlab utility. When saving a data file over the network, Maestro first saves it to a temporary shadow file on the local disk. Until this change, the shadow filename was different from the remote filename because the letters ".TMP" were appended (eg, target filename = "mytrial.0002"; shadow filename = "mytrial.0002.TMP"). But this meant that the data filename seen by the Plexon was incorrect (it included the ".TMP" suffix). The ".TMP" string is no longer appended.


Version 2.0.1, released 21 Jun 2006

    • Modified definition of the XYScope target Rectangular Annulus to include X,Y offsets (relative to target center) for the rectangular hole within the annulus. These default to (0,0), corresponding to a concentric hole. Unfortunately, this small change impacted the format of the Maestro experiment document and the data file. Analysis programs readcxdata() and XWork have been updated accordingly.

    • Minor change in error reporting to distinguish between RMVideo "skipped frame" and "duplicate frame" error conditions.


Version 2.0.0, released 11 Jun 2006

    • RMVideo. The old VSG2/4 framebuffer video card was replaced by an entirely new framebuffer video solution: Remote Maestro Video, or RMVideo for short. RMVideo is an OpenGL application running on a separate Linux workstation, communicating with Maestro over a private point-to-point Ethernet link. RMVideo supports all of the old VSG2/4 targets, and it implements XYScope-like targets as well. This new video platform allows us to take advantage of today's modern graphics cards, offering far superior performance to the VSG. It can handle drifting Gabor targets and supports protocols that involve runtime adjustments in target trajectories -- such as velocity stabilization. The only disadvantage is that it requires a second computer. Numerous changes to the Target Editor to support defining RMVideo targets.

    • Velocity stabilization in Trial mode can now be engaged on a per-target and per-segment basis. The Trial Editor has been modified accordingly: controls that specified the open-loop segment, etc have been removed from the Special Features group, and an additional row has been added in the segment table for each target so that the user can specify the velocity stabilization mode (OFF, H+V, H ONLY, and V ONLY) and "snap to eye" flag for each target during each segment of the trial.

    • Analysis program XWork and Matlab MEX function readcxdata() have been updated to handle data files generated by this version.

    • Added some informational readouts to the Video Display tab.

    • Added trial header flags to override the application of the global target position/velocity transforms specified on the Other Params tab of the Trial mode control panel. These flags are exposed by four new check boxes in the General section on the Trial Editor.

    • Bug fixed: Have observed that the dependency-locking scheme -- which prevents deletion of dependent targets, perturbations, or channel configurations -- may break. As a result, one may be unable to delete an object because its lock count is wrong. Fixed the problem by providing a workaround. All object lock counts are now recomputed fresh when the document is read from file. If, in the course of editing an experiment document, you find that you cannot delete an object, save the document and read it back in. If you still cannot delete the object, then most likely there's a trial or stimulus run definition that depends upon it!