Trials

A Maestro trial is a flexible, full-featured experimental protocol in which up to 25 different moving targets are presented while the subject's behavioral and neuronal responses are recorded and saved in a data file. Target trajectories are defined over a series of trial segments, with as many as 30 different segments allowed in a single trial. Analog data is recorded at 1KHz, digital events like spike arrival times are timestamped with 10 µs accuracy, and the position of the animal Chair is updated once per millisecond. Marker pulses may be delivered at the start of each trial segment. Trials support animation of either XYScope targets with a programmable refresh rate as high as 500Hz, or RMVideo framebuffer video targets at a fixed rate (the vertical refresh of the display). Selected components of a target's segmented trajectory may be modulated by a periodic or noisy perturbation, and targets can be stabilized on the retina on a segment-by-segment basis. In addition, a number of specialized features have been implemented in Trial mode to support two-choice learning experiments or target search tasks.

As of Version 3.3.0, a trial may define up to 10 different random variables which, in turn, can govern the duration of a segment, a target's window position, velocity or acceleration during a segment, and an extended target's pattern velocity or acceleration during a segment. A random variable changes value with each presentation of the trial, IAW a defined distribution -- uniform, normal (gaussian), exponential, or gamma. You can also define a random variable as a function of other random variables.

In a typical "experiment" in Trial mode, the subject responds to a sequence of trials drawn randomly or otherwise from a selected trial set. Individual trial definitions are viewed and edited on the Trial Editor, one of the panels in the tabbed window occupying the right-hand side of the client area in Maestro's frame window. The trial is by far the most complex object in Maestro, and that is reflected in the layout of the Trial Editor. The trial's segments and participating targets, including per-segment trajectory parameter, are displayed and edited in the Segment Table, which occupies much of the editor panel. Above that table is a tabbed pane in which all the other widgets are organized: the most-used parameters are located on the Main tab; the perturbations table and PSGM parameters are found on the Perturbations/PSGM tab; and random variables are defined on the Random Variables tab.

The following sections describe a trial's features and defining parameters, and how to use the Trial Editor to modify the trial's definition.

• • General Parameters: Channel configuration for the trial, record start segment, failsafe segment, etcetera.

  • Special Features: Two-choice protocols, search task, mid-trial rewards, staircase trials for psychophysics experiments, and other specialized operations.

  • Pulse Stimulus Generator Module (PSGM): Parameters defining the presentation of a PSGM pulse train during the trial. The PSGM delivers current pulses into the brain tissue via an extracellular electrode.

  • Perturbations Table: Defines any perturbations presented during the trial, specifying the perturbation object and its amplitude, the target affected, the segment in which the perturbation begins, and the trajectory component that is modulated by the perturbation.

  • Random Variables Table: A list of 10 random variables, any subset of which may be defined for use and then assigned to various parameters in the segment table.

  • Segment Table: Defines the trial's segments and the list of participating targets. Target window position, velocity and acceleration can be specified on a per-segment basis.

If you read through the section in its entirety, you'll likely conclude that defining a trial is time-consuming and tedious. In practice, it is often the case that the trials in a given trial set will be very similar in structure: same number of segments, same participating targets, similar segment durations, and so on. Thus, once you've created one trial, simply copy it as many times as you need, then load each copied trial into the Trial Editor, and make any necessary changes to its definition. In addition, if you are proficient at writing Matlab scripts, it is now possible to generate an entire Maestro experiment document in Matlab, saving the content in a plain-text format that can then be imported into Maestro. For more information on this subject, see Scripting Experiments in Matlab.