How to correct millisecond timing errors in your studies
Find out more
To improve replication and enhance credibility researchers should self-validate, or self-certify, their own studies in terms of millisecond presentation, synchronization and response timing accuracy.

Currently self-validation of millisecond timing accuracy can only be done quickly and easily with a Black Box ToolKit. This acts as a programmable virtual human that can detect and respond to stimulus events with sub-millisecond accuracy. It enables you, the researcher, to check the accuracy of your own paradigm whilst running in-situ on your own equipment.


  1. Set-up – Hook up external stimulus sensors (opto-detectors, mics, TTL) and a response device (response pad, robotic key actuator, sounder, TTL)
  2. Test – Use a wizard to select a stimulus pattern to automatically respond to/event mark. Then choose the exact response time, response device and duration
  3. Analyze – Check your stimulus, response and synchronization timings across up to 36 channels. Compare what your experiment recorded with what the BBTK saw

Until the advent of the Black Box ToolKit checking the timing of traditional computer-based experiments was extremely difficult, time consuming and required highly specialist equipment and knowledge.

The Black Box ToolKit on the other hand allows you, the researcher, to quickly and easily self-validate the timing accuracy your own paradigm whilst running unmodified on your own equipment. In essence the Black Box ToolKit acts as a highly accurate "virtual human" which you tell to make responses to as a result of various stimuli it detects. This low cost kit can be used to detect where timing errors occur and to help you make corrections if needed.

The ToolKit can monitor for visual events, auditory presentations, TTL event markers or sync pulses from EEG and MRI scanners and can generate responses via tones (vocal), TTL, Active Switch Closure (key down) or robotic key actuator. What's more it does all this with sub-millisecond accuracy.

The Black Box ToolKit offers a wide range of inputs and outputs that helps you interface with your own equipment quickly and easily with little or no modification. The new Black Box ToolKit version 2 offers even more interfacing options and is just as easy to use. To see how easy, check out the worked example below or watch the demo video of the new BBTK v2 in action.


The Black Box ToolKit version 2 (Elite model shown).
The BBTK v2 offers more interfaces and a new easy to use software suite.

How to self-validate a simple visual-stimulus response experiment
Lets say you were using a leading experiment generator to test simple visual reaction time. Your equipment might consist of a standard Microsoft Windows PC, a TFT monitor for presenting stimuli and a USB keyboard for registering responses. The experiment itself might display an image for 100 milliseconds and then wait for a response (space bar) before displaying the next trial after a 1,000 millisecond ISI.

Despite this being a simple experiment there are numerous areas where timing errors can occur. Here we're just going to focus on three.

Questions we should ask:

  1. Will the TFT display the image in a timely fashion due to input lag?
  2. Does the experiment generator actually display each image for 100 milliseconds?
  3. Is the reaction time might be wrong?

To begin with we'd hook up an opto-detector to the CRT or TFT monitor we were using. Next we'd tack an Active Switch Closure to the space bar of our keyboard so that we could trigger a response (alternatively we could use the BBTK robotic finger actuator). That takes care of monitoring stimuli and generating responses.

Hooking up an opto-detector to a CRT or TFT monitor.
A BBTK v1 shown but the new v2 operates on the same principle.

Starting the BBTK v2 software

Defining what stimuli to respond to is child's play as the screen gab of the actual BBTK software shows. First say how many responses you want to make (A) then click on the green stimulus lines you want to respond to (B). Next click on the red responses you want to make (C). Finally define the characteristics of the response. In this case a Reaction Time of 300 milliseconds (D) and a key down duration of 100 milliseconds (E). Then say how long you want to keep monitoring for (F).

Defining what stimulus to respond to and how to respond to it in the Black Box ToolKit version 2 DSCAR software.

We're now set to self-validate the timing accuracy of our paradigm. To do this we'd start the Black Box ToolKit and the experiment and let the BBTK respond
to each stimulus as they are presented.

Question 1: Was the image presented in a timely fashion?

If we look at Black Box ToolKit Digital Oscilloscope (part of our data analysis software) the answer is no. Using the two measurement cursors we can see that the ISI was 1,363 milliseconds where as we would expect it to be 1,300 milliseconds. That is, 1,000 millisecond gab between trials, 300 milliseconds for a Reaction Time. So the input lag of the TFT is around 64 milliseconds.

Trace (A) shows the stimulus images being displayed. Trace (B) shows the Black Box ToolKit generating responses to register a Reaction Time. The measurement cursors (C) are used to measure the time in milliseconds between any two points. The ISI is read off as (D).

The 20 channel digital oscilloscope of the Black Box ToolKit v2 software.

Question 2: Were stimulus images displayed for 100 milliseconds?

To answer this question we could either use the measurement cursors or use the "Line by Line Analysis" spreadsheet which is shown along with the digital scope. Here we can see that images are displayed for 27 milliseconds too long (A).

Simple to understand readings from sensors and response devices.

Question 3: Was the Reaction Time recorded by our experiment accurate?

To determine this we would open the data file recorded by our experiment generator and compare it with the response we generated in response to the visual stimulus. Below we can see that due to polling loops in our keyboard the Reaction Time was not 300 milliseconds but that it varied considerably (A).

Comparing Reaction Times recorded by your own software to
that generated by the Black Box Toolkit.

If you we using more sensors or response generation interfaces, timings for those would appear on other lines of the scope and other columns of the spreadsheet. The same applies for TTL event markers for EEG and sync pulses in fMRI studies.

Self-validating your experiment timing really is as simple as that!