Assgn 8: Delay and Cancellation of Saccades

Background:  Despite the fact that saccades are the fastest eye movements (400 deg/sec) they are the slowest to get started, with basic latencies ranging from 180 - 380 msec. Here you will model a version of Larry Stark's sampled data model to account for the variability of saccade latency.

See the artwork below, from the archive website. You can use your work on Assgn 7 as the saccade generator which the delay system feeds into. You will want to include a data sample-and-hold circuit after the visual delay and cancellation cutout. In the "Motor" block should be a logic test and a means of saccade cancellation.

Requirements:
(1) Generate the A + B Wheeless stimulus, seen above, and in Fig. 4.16-18, where W = 100 ms. Attending to Fig 4.20, Generate a 5Hz sample clock. The sample clock can go into the sample-and-hold you designed into Asgn 7. You can vary when the a sample pulse first encounters A or B stimulus by incrementing or decrementing the "phase delay" in the pulse generator, with units in seconds, not degrees. What would be a good value for width, or duty cycle, of the sample pulse?

(2) Sampled Data: Have a 5Hz low-duty-cycle (4%?) pulse train drive a sample-and-hold clock. Design a cancellation circuit, with possible inputs from "Assgn 7 hold", current-visual and Motor Delay. Output goes to the OMN. See suggestions below:

The Motor Delay should be 50 msec, and the Visual Delay 130 msec. The saccade generator should be your circuit from Assgn 7. (Thus there may be 2 sample-and-hold circuits in your model!)

(3) Cancellation: At the end of the 50 msec motor delay the system needs to decide whether to cancel the forthcoming saccade. What is the criterion for cancellation?

(4) Testing total saccade delay w.r.t. data in Fig 4.19, 4.21: There are two possible outcomes from a subject looking at the A + B stimulus: either A + B saccades, or B-only.
Consider A+B. Your model should have a range of 180 - 230 msec for A+B delay, depending on the phase relationship between the sample clock and the A portion of the stimulus.
As you can see in Fig 4.19, the minimum time for a B-only saccade is 380 msec. It appears that this duration includes the 100 msec that A lasts. What latency does you model generate when the sample clock falls at 101 msec after start of A? (1 msec after A ends). Attend to this sentence from RHSC2, p. 91, "It is also necessary to assume the cancel signal lasts at least 50 ms, as otherwise the earliest type B responses, due to sampling occurring just after the second target displacement, would not be cancelled." Can you incorporate some version of that sentence in your model to "correct" the early responses to type B responses?

Can you demonstrate the longest latency, 530 msec, of Fig 4.19, with your model?

Possible FTQ: What happens to A+B responses if there is no cancellation? What is a physiological basis for the 5Hz sample clock?

Reading: You have available chpt 4 of RHSC2. Study Fig's 4.16-21.

Free Advice: Read about Robinson's no-sample-clock method of explaining saccade latency, top of p. 93.