Fig. 6: Investigation of the relationship between WM load and the time spent in states 1 and 3 in the 4-subject dataset.

a The portion of time spent in states 1 and 3, averaged across sessions, subjects, and tasks. The shaded areas mark stimulus presentation periods, and the non-shaded areas mark the delay periods. b The portion of time spent in states 1 and 3 during the verbal recognition task, averaged across subjects and sessions. c The verbal recognition task used as a control. Letters are presented verbally in sequential order, and at the end of the presentation, participants responded ‘yes’ if the letter Q had been presented and ‘no’ otherwise. d The time spent in states 1 and 3 during the WM-Grid task. Each dot represents the mean time for each session. Above: the time spent in state 3 during the delay periods. Load 0 is the 1000 ms before presentation of the first item, and load 5 is the 1000 ms delay period after the presentation of the fifth item. Below: the time spent in state 1 during the 300 ms presentations of each stimulus 1–5. e The time spent in states 1 and 3 during the Odd One Out task. All analysis and time periods are identical to those of WM-Grid (apart from the Odd One Out task not being included for measurement session 3). Source data are provided as a Source Data file.