Montreal Imaging Stress Task

Script Author: Katja Borchert, Ph.D. (katjab@millisecond.com), Millisecond

Created: January 16, 2015

Last Modified: January 18, 2025 by K. Borchert (katjab@millisecond.com), Millisecond

Script Copyright © Millisecond Software, LLC

Background

This script implements an Inquisit version of the Montreal Imaging Stress Test as described in:

Dedovic, K.; Renwick, R,; Khalili Mahani, N.; Engert, V.; Lupien, S.J. & Pruessner, J.C. (2005). The Montreal Imaging Stress Task: using functional imaging to investigate the effects of perceiving and processing psychosocial stress in the human brain. J Psychiatry Neurosci, 30, 319-325.

The Montreal Stress Task was developed as a test to be used under imaging conditions (e.g. fMRI, PET) where participants might have limited access to keyboards/touchscreens. This script is a best guess effort of Millisecond based on published information.

Duration

40 minutes

Description

Participants are asked to solve arithmetic problems of 5 different difficulty levels under 3 different stress conditions. In the high stress condition (experimental condition), the participant's performance is manipulated to be relatively low by a) timing performance (e.g. if too many are correct, less time is allocated) b) comparing the participant's performance to the "average" performance (80-90% correct) by means of a "performance bar" In the control condition, the participants are asked to solve the same problems but without overt timing and without presenting the performance bar. In the rest condition, participants are simply asked to rest. In all conditions the participants are asked to control the selection of the solution by "dialing" via mouse selections: left button (dials counterclockwise), right button (dials clockwise), middle button (submits the response).

Procedure

(A) 5 levels of difficulty: tested in blocked format

In this script version, the levels are defined as follows:
1: 2 1-digit integers (0-9), only + or -, solution 0-9 (1 digit)
2: 3 1 digit integers (0-9), only +,- (no repeated operations, order of operations random)
3: 3 integers, 1-2 integers are double digits (0-99), +,-,* (no repeated operations, selection and order of operations random), solution 0-9 (1 digit)
4: 4 integers, 1-2 integers are double digits (0-99), +,-,* (no repeated operations, selection and order of operations random), solution 0-9 (1 digit)
5: 4 integers, 1-4 integers are double digits (0-99), *, / are a must, then either + or - (order of operations random), solution 0-9 (1 digit)

the problems are pregenerated for this script (see also Dedovic et al, p.320)

(B) 4 different conditions:
1) Training:
- administered outside of imaging equipment
- uses the mouse to control the selection of the solution (the 'dials') -> Dedovic et al (2005) use keyboard input for training
- minimum suggested duration: 2 minutes for entire training block (this translates to Min = 24s for each level of difficulty);
he default in this script is set to 5 minutes (=1 minute/level of difficulty)
- presents a random order of the 5 difficulty level blocks
- calculates average time participants used to respond to each problem (regardless of accuracy) for each level of difficulty.
The uncorrected latencies are used in this script to ensure that initial timelimits can be calculated.
- does not put a time limit on performance
- does not give overall performance feedback in the form of a performance bar
- presents feedback (correct, incorrect) for individual problem for 500ms (editable parameter)
training ends with a slide that says "Please Wait" (to continue the experimenter has to press the Spacebar - participant cannot move from here with the mouse)

2) Experimental:
- minimum suggested duration: 2 minutes per level of difficulty
this timelimit is based on our interpretation of (->"individual runs" p.321, right column)
- presents a random order of the 5 difficulty level blocks
- puts a time limit on performance
a) at the start of a block uses training mean latency for the currently tested level of difficulty but shortens it by 10% (= 90% of the training mean response times)
b) continuously tracks the performance and response time of the last 3 trials
=> if the last 3 trials are all correct: adjust time limit by using the average response duration of the last 3 trials but shorten it by 10%
=> if the last 3 trials are all errors (or timeouts): adjust time limit by lengthening the current timeout by 10%
c) presents a timer on screen that counts down the seconds (Dedovic et al, 2005 used a progress bar)

- presents a performance bar that presents the "average" performance as being in the green (good) region of performance and the participant's
performance as being (likely) in the red (bad) region of performance (average displayed performance is calcuated across experimental difficulty levels)

- presents feedback (correct, incorrect, timeout) for individual problems for 500ms (editable parameter)

- adds an intertrial interval (ITI): the duration of the intertrial interval is based on parameters.iti_test (editable parameter) which gets
adjusted by the difference of the average response latency for the tested level (based on training performance) minus the current response latency.
If the current response latency was faster, the ITI increases. If the current response latency was slower (unlikely as a timeout is imposed) the ITI decreases
the ITI was added so that control and experimental trials could (roughly) be matched in frequency.
(Dedovic et al, 2005, p.321: "To match the frequency of mental arithmetic tasks [in experimental and control condition], the time between tasks is varied as a function of the time limit
imposed during the experimental condition, so that the total number of tasks presented per condition is identical.")
The idea is that if the response time is equal to the training response time (which is run under similar conditions to the control condition)
the ITI is parameters.iti_test. If it's shorter (because of the timeout) then the ITI gets adjusted up - otherwise down.
This works theoretically AS LONG AS there is still a positive ITI (>=0ms) left at the end. If a control trial (or experimental trial, though that is less likely)
is longer than the average response time by parameters.ITI then the number of trials in the experimental and control condition might differ.

each experimental blocks ends with a slide that informs participants of their average performance; followed by a slide that says "Please Wait"
(to move on the experimenter has to press the Spacebar - participants cannot move on from here with the mouse)

3) Control:

- minimum suggested duration: 2 minutes per level of difficulty
- presents a random order of the 5 difficulty level blocks
- does not put a time limit on performance
- does not present a performance bar
- presents feedback (correct, incorrect, timeout) for individual problems for 500ms (editable parameter)
- adds an intertrial interval (ITI): the duration of the intertrial interval is based on parameters.iti_test (editable parameter) which gets
adjusted by the difference of the average response latency for the tested level (based on training performance) minus the current response latency.
If the current response latency was faster, the ITI increases. If the current response latency was slower (unlikely as a timeout is imposed) the ITI decreases
the ITI was added so that control and experimental trials could (roughly) be matched in frequency.
(Dedovic et al, 2005, p.321: "To match the frequency of mental arithmetic tasks [in experimental and control condition], the time between tasks is varied as a function of the time limit
imposed during the experimental condition, so that the total number of tasks presented per condition is identical.")

each control blocks ends with a slide that says "Please Wait"

4) Rest: inside imaging equipment

- presents the userinterface with a message to take a break and not move the mouse until told to do so
the rest block ends with a slide that says "Please Wait" (to move on the experimenter has to press the Spacebar - participant cannot move from here with the mouse)

The order of the 3 test conditions (experimental, control, rest) is counterbalanced by groupnumber (6 different groupnumbers run the 6 possible orders)
To change the experimental procedure (e.g. if no rest condition should be run), go to section EXPERIMENT and delete the blocks
that should not run.

groupnumber1: exp, ctrl, rest
groupnumber2: exp, rest, ctrl
groupnumber3: ctrl, exp, rest
groupnumber4: ctrl, rest, exp
groupnumber5: rest, exp, ctrl
groupnumber6: rest, ctrl, exp

Stimuli

this script uses pregenerated sequences for the 5 levels of difficulty. The problems used can be edited under
section "Editable Lists"

Instructions

this script uses instructions that are not original to Dedovic et al (2005).
They can be edited in script montrealstresstest_instructions_inc.iqjs (language-dependent)

Summary Data

File Name: montrealstresstest_summary*.iqdat

Data Fields

Raw Data

File Name: montrealstresstest_raw*.iqdat

Data Fields

Parameters

The procedure can be adjusted by setting the following parameters.