Millisecond Forums » Millisecond Forums » Inquisit 6 » Spatial Multi-Arrangement task

Spatial Multi-Arrangement task

Xiaoping — Tue, 09 May 2023 12:41:43 GMT

Hello,

I would like to know if it is possible to script multiple arrangement tasks using Inquisit. The task is to efficiently obtain pairwise similarity of a set of items (words, images, etc.) by arranging items in a 2D space (Kriegeskorte & Mur, 2012). In the initial trial, all of the items will be arranged according to the similarity between items (close in distance means high similarity), and in the following trials, subsets of items will be selected for more arrangement to better estimate pairwise similarity ("zoom in"). The (dis)similarity matrix derived from the task can be used to reveal mental representations of items and even be associated with other matrices derived from neuroimaging data.

The task requires participants to click and drag items and I see Inquisit can do this. Additionally, there are also algorithms to determine which subset of items to be arranged in the next trial based on participants' performance. I have trouble finding this type of task in the test library but am worried if I missed any. If this has not been developed yet, I was wondering if there is any chance this will be developed, as I see this task is of interest to many researchers who are using representational similarity analysis (RSA).

Thank you very much!
Xiaoping

Kriegeskorte, N., & Mur, M. (2012). Inverse MDS: Inferring dissimilarity structure from multiple item arrangements. Frontiers in psychology, 3, 245.

RE: Spatial Multi-Arrangement task

Xiaoping — Tue, 09 May 2023 12:41:43 GMT

+x[quote]

[b]Dave - 5/9/2023[/b]
+x[quote]

[b]Xiaoping - 5/9/2023[/b]
Hello,

I would like to know if it is possible to script multiple arrangement tasks using Inquisit. The task is to efficiently obtain pairwise similarity of a set of items (words, images, etc.) by arranging items in a 2D space (Kriegeskorte & Mur, 2012). In the initial trial, all of the items will be arranged according to the similarity between items (close in distance means high similarity), and in the following trials, subsets of items will be selected for more arrangement to better estimate pairwise similarity ("zoom in"). The (dis)similarity matrix derived from the task can be used to reveal mental representations of items and even be associated with other matrices derived from neuroimaging data.

The task requires participants to click and drag items and I see Inquisit can do this. Additionally, there are also algorithms to determine which subset of items to be arranged in the next trial based on participants' performance. I have trouble finding this type of task in the test library but am worried if I missed any. If this has not been developed yet, I was wondering if there is any chance this will be developed, as I see this task is of interest to many researchers who are using representational similarity analysis (RSA).

Thank you very much!
Xiaoping

Kriegeskorte, N., & Mur, M. (2012). Inverse MDS: Inferring dissimilarity structure from multiple item arrangements. Frontiers in psychology, 3, 245.
[/quote]

Having taken a look, I don't think the required math is reasonably feasible with Inquisit's current capabilities, specifically its present lack of multidimensional arrays and matrix algebra functionality. A "toy" implementation (very small item set) might be possible, but with realistic set sizes, the lack of certain capabilities unfortunately & quickly becomes a dealbreaker.
[/quote]

Understand. Thank you, Dave!

RE: Spatial Multi-Arrangement task

Dave — Tue, 09 May 2023 12:19:12 GMT

+x[quote]

[b]Xiaoping - 5/9/2023[/b]
Hello,

I would like to know if it is possible to script multiple arrangement tasks using Inquisit. The task is to efficiently obtain pairwise similarity of a set of items (words, images, etc.) by arranging items in a 2D space (Kriegeskorte & Mur, 2012). In the initial trial, all of the items will be arranged according to the similarity between items (close in distance means high similarity), and in the following trials, subsets of items will be selected for more arrangement to better estimate pairwise similarity ("zoom in"). The (dis)similarity matrix derived from the task can be used to reveal mental representations of items and even be associated with other matrices derived from neuroimaging data.

The task requires participants to click and drag items and I see Inquisit can do this. Additionally, there are also algorithms to determine which subset of items to be arranged in the next trial based on participants' performance. I have trouble finding this type of task in the test library but am worried if I missed any. If this has not been developed yet, I was wondering if there is any chance this will be developed, as I see this task is of interest to many researchers who are using representational similarity analysis (RSA).

Thank you very much!
Xiaoping

Kriegeskorte, N., & Mur, M. (2012). Inverse MDS: Inferring dissimilarity structure from multiple item arrangements. Frontiers in psychology, 3, 245.
[/quote]

Having taken a look, I don't think the required math is reasonably feasible with Inquisit's current capabilities, specifically its present lack of multidimensional arrays and matrix algebra functionality. A "toy" implementation (very small item set) might be possible, but with realistic set sizes, the lack of certain capabilities unfortunately & quickly becomes a dealbreaker.