1 of 22

Experiment control with computers

Attila Krajcsi

​

​

​

​

Last update: 2025 February

​

2 of 22

Outline

  • General problems
  • Some specific systems
  • Some general building blocks

3 of 22

Experiment control with computers

General Problems

4 of 22

Precise timing

  • Millisecond precision timing
    • Previously: was needed
    • Today: is enough
  • Hardware precision/capability
    • Usually not an issue anymore
  • Operating system interruption
    • Other programs should not run
  • Browser-based studies
    • In general, it requires better hardware and better software solution

5 of 22

Precise stimulus presentation

Monitor

  • Screen refreshing frequency
    • Limits the shortest time to present a stimulus
    • Do you need a more expensive monitor?
  • Time of refresh
  • Software layers
    • DirectX, OpenGL, Vulkan, etc.
    • Which operating systems are supported (Mac, Android, Linux, web-based)?

6 of 22

Precise timing of response measure

  • Keyboard
    • PS/2, USB or notebook
  • Response box
    • Fast signal (e.g. parallel port) - hardly available anymore
    • Or timestamp
  • It’s (mainly) not the question of the speed but the delay and noise (variance) in the delay

7 of 22

Synchronization with other tools

  • EEG
  • Eye movement (and access to the eye movement data)
  • fMRI (fMRI starts the experiment control software)

​

  • Frequently used ports:
    • USB
    • Ethernet port
    • Still, you might find some parallel port or game port

​

8 of 22

Interactivity

  • Stimuli depend on responses, e.g.,
    • Feedback
    • Learning reaches a criterion
    • Rhythm on your own pace

9 of 22

Some other differences among experiment control solutions

10 of 22

Graphical user interface vs. programming

  • Graphical user interface
    • Easy to learn (e.g., a few hours)
    • Most frequent solutions are available
  • Programming
    • You should understand computer programs
    • Harder to learn (basic things can be learned in a week)
    • More flexible
    • Surprisingly, it can be easier to overview

​

11 of 22

Operating systems

  • Operating systems
    • Windows
    • Mac
    • Linux
    • Android
    • Browser-based

​

  • Usually
    • Win only
    • Cross platform
      • Android support is getting better
    • (Mac only)
    • Web-based

​

​

12 of 22

Price

  • Free
    • Open source/Closed source
  • Proprietary
    • The software vs. the environment (e.g., Matlab, Windows)
    • Price for a period or one time price
    • Only for one computer or more flexible
      • Editing only on one computer
    • Group and student discounts
    • Etc.

13 of 22

A few examples

14 of 22

A few examples

  • PXLab
  • PsyScope
  • Inquisit

​

  • SuperLab
  • Vision Egg
  • EXPE

​

​

  • Softwares for Psychophysics

15 of 22

Open Sesame

16 of 22

PsychoPy

17 of 22

Some general building blocks

18 of 22

A trial

  • Trials are series of events
  • Events are mostly pictures and texts on the screen, sounds, videos, and responses from the participant

​

​

19 of 22

Building blocks of the events

  • Objects (coming from object oriented programming)
    • E.g., text, sound, picture, etc.
  • Properties of objects (coming from object oriented programming)
    • E.g., color, size, position on the screen, etc.

20 of 22

Trials organized in loops

  • Trials repeated
  • Built-in loops, e.g., blocks

21 of 22

Changing parts of the trials

  • Some properties change trial by trial or block by block, some others don’t
  • Some properties change in trials, some others in blocks, etc.

22 of 22

Summary

  • Special solutions for experiment control
    • Stimulus presentation
    • Response measurement
    • Connecting devices
    • Changing the experiment based on online performance
  • How to choose a software (beyond the previous viewpoints)
    • OS/environment support
    • Price
    • Flexibility