Sequence learning
Test whether responses on a task with four positions speed up with practice, and what underlies any change.
Method
Procedure
- A circle fills in one of four positions. Press the key for that position as quickly as you can without making mistakes.
- There are five blocks of 50 trials, with a short rest between blocks.
- If you press the wrong key, the circle stays filled until you press the right one (for up to 3 seconds).
- At the end there is one question and a short final task.
Responses
- D, F, J, K
- positions 1 to 4, from left to right
- Touch
- tap the circle, or the buttons 1 to 4
Measures
The reaction time on each trial and whether the first key press was correct.
Practice
Eight practice trials with positions in a random order tell you after each trial whether you pressed the right key and how fast you were.
Knowing the design in advance would change the result, so it is explained with the results and the background section stays closed.
Terms
- Reaction time: The time from the circle filling to your key press, in milliseconds.
- Block: A run of trials done without a break.
- Implicit learning: Learning that shows in performance without the person necessarily being able to describe what was learned.
- Basal ganglia: Structures deep in the brain that work with the cortex in selecting and learning movements.
Experiment
Responses stay in this browser tab. Nothing is stored or sent, and everything is lost when the page is closed or reloaded.
Results
Interpretation
Blocks 1 to 3 and 5 followed a repeating ten-position sequence; block 4 used random positions. Speeding up over the first blocks can come from general practice, so the key comparison is block 4 against blocks 3 and 5 on either side of it. Slower responses in block 4 indicate that you had learned something about the order. The reshuffling test gives a percentage: the smaller it is, the less likely it is that a slowing as large as yours would arise by chance (below 5% is the usual convention). The final question and task show how much of the sequence you could report.
Comparison with published studies
A clear slowing in the random block follows the pattern reported by Nissen and Bullemer (1987). The paper gives no single typical value for comparison.
Limitations
One random block of 50 trials is compared with two sequence blocks, so a single slow spell, for example a lapse of concentration in block 4, can create or hide an effect. The questions at the end are a rough check of awareness, not a full test.
The result comes from one person in one session, on a device whose timing and display are not calibrated. It describes this run, not a stable characteristic of the participant, and a second run could differ by chance alone. Research studies test many participants under controlled conditions and report the average with its variability. The experiment is educational; it is not a test of ability or health.
One row per trial: block, block type (sequence or random), trial, position, your first key press, reaction time, whether the first press was correct, missed or too fast; then one row for each position you entered when reproducing the sequence, with your answer to the question. In the file, 1 means yes and 0 means no. The file is created on this device and can be opened in a spreadsheet.
Background and references (to be read after taking part)
Background
Original findings
In the serial reaction time task introduced by Nissen and Bullemer (1987), the target follows a repeating ten-item sequence of positions without the participant being told. Nissen and Bullemer (1987) found that performance improved more with the repeating sequence than with random positions, and that learning was impaired when attention was occupied by a secondary task. Here the fourth block is random: each position appears as often as in the sequence, and, as in the sequence, never twice in succession and never as a back-and-forth reversal, so the blocks differ only in the order of the positions. Slowing in the random block relative to the sequence blocks either side of it (blocks 3 and 5) is taken as a measure of knowledge of the order, separate from general practice and fatigue; with a single session it is tested against chance by reshuffling the reaction times between the three blocks. Learning of this kind can appear in reaction times in people who report little awareness of the sequence, although how far it is truly implicit is still debated; the questions at the end give only a rough check.
Neural basis
Sequence learning of this kind is thought to depend on circuits linking the basal ganglia, in particular the striatum, with motor and premotor areas of the cortex, and on the cerebellum. Some studies of people with conditions affecting the basal ganglia, such as Parkinson’s disease, have reported reduced learning in this task. The hippocampus and prefrontal cortex have been suggested to contribute more when people become aware of the sequence.
References
- Nissen, M. J., & Bullemer, P. (1987). Attentional requirements of learning: evidence from performance measures. Cognitive Psychology, 19(1), 1–32. https://doi.org/10.1016/0010-0285(87)90002-8
Last updated