07-14-2026, 12:43 PM
Dear Team,
I'm trying to put together a digit span task but I think I will need your help to implement it fully as there are some features that your platform does not have (yet). I would like to make two versions of the digit span task, a forward and backward version. The forward version is the simplest. Participants must recall digits in the order in which they were displayed. For example, if they see 7 3 6 9 then they type 7 3 6 9. Regarding the backward version, they must recall the digits in their opposite order of presentation. Using the example given, they'd have to respond with 9 6 3 7.
The forward and backward versions both commence with a trial of 2 digits in length. The forward version has a max of 9 digits and the backward task a max of 8. A trial of length n can be attempted at most twice. If a participant answers correctly on the second attempt they proceed to the next trial, whose length is now +1 relative to the previous trial. If both attempts are incorrect the task stops entirely. Similarly, the tasks also stops when a participant reaches the max number of digits, i.e. 9 for the forward and 8 for the backward.
Trial layout:
- fixation cross for 2000ms
- individual digits each displayed for 1000ms
- response box
- fixation cross for 2000ms
- etc.
For the above layout, there are some features on FindingFive which offer good starting points but they don't quite hit everything on the head. Regarding the procedure, there is the existing node "branching": { } which I think holds some promise because of the "min_score", "branches" and "iterations" functions but I wonder if we could employ something similar but in the Trial Templates?
- A "min_score": TRUE --> would be required in order to progress onto the next trial.
- "iterations": {"max": 2} --> would ensure each trail can be attempted at most twice before "branches" comes into play.
- "branches": {"original_trial_template": TRUE, "Exit": FALSE } --> I'm wondering if we could create a loop here within the trial_template that the "branching": {} is attached. Whereby, if TRUE then a participant is looped back to the trial template but this time it has +1 digit, else they are directed to the "Exit" Trial Template. This trial_template would also need to be specified with a max of 8 or 9 digits, depending on the version.
Another thing we'd need to look out for, each trial will generate n digits and in order to either a) progress to the next trial, b) try once again or c) stop the task, responses will need to be compared against the randomised digits previously presented. However, since the digits are randomised a function such as "target" within a response will not be fruitful. Is there a way to set up maybe a TRUE/FALSE value within a response that can compare a response given against stimuli/digits previously presented?
I hope these thoughts don't cause too much of a headache but I'm very happy to clarify and develop these ideas further.
Cheers,
Mary-Kate
I'm trying to put together a digit span task but I think I will need your help to implement it fully as there are some features that your platform does not have (yet). I would like to make two versions of the digit span task, a forward and backward version. The forward version is the simplest. Participants must recall digits in the order in which they were displayed. For example, if they see 7 3 6 9 then they type 7 3 6 9. Regarding the backward version, they must recall the digits in their opposite order of presentation. Using the example given, they'd have to respond with 9 6 3 7.
The forward and backward versions both commence with a trial of 2 digits in length. The forward version has a max of 9 digits and the backward task a max of 8. A trial of length n can be attempted at most twice. If a participant answers correctly on the second attempt they proceed to the next trial, whose length is now +1 relative to the previous trial. If both attempts are incorrect the task stops entirely. Similarly, the tasks also stops when a participant reaches the max number of digits, i.e. 9 for the forward and 8 for the backward.
Trial layout:
- fixation cross for 2000ms
- individual digits each displayed for 1000ms
- response box
- fixation cross for 2000ms
- etc.
For the above layout, there are some features on FindingFive which offer good starting points but they don't quite hit everything on the head. Regarding the procedure, there is the existing node "branching": { } which I think holds some promise because of the "min_score", "branches" and "iterations" functions but I wonder if we could employ something similar but in the Trial Templates?
- A "min_score": TRUE --> would be required in order to progress onto the next trial.
- "iterations": {"max": 2} --> would ensure each trail can be attempted at most twice before "branches" comes into play.
- "branches": {"original_trial_template": TRUE, "Exit": FALSE } --> I'm wondering if we could create a loop here within the trial_template that the "branching": {} is attached. Whereby, if TRUE then a participant is looped back to the trial template but this time it has +1 digit, else they are directed to the "Exit" Trial Template. This trial_template would also need to be specified with a max of 8 or 9 digits, depending on the version.
Another thing we'd need to look out for, each trial will generate n digits and in order to either a) progress to the next trial, b) try once again or c) stop the task, responses will need to be compared against the randomised digits previously presented. However, since the digits are randomised a function such as "target" within a response will not be fruitful. Is there a way to set up maybe a TRUE/FALSE value within a response that can compare a response given against stimuli/digits previously presented?
I hope these thoughts don't cause too much of a headache but I'm very happy to clarify and develop these ideas further.
Cheers,
Mary-Kate
