The tested treatment is a combination of neuropsychological rehabilitation procedures: learning, episodic memory recall after a delay, selective attention, inhibition of predominant responses and awareness of deficits.

Non-confabulators (brain injured patients but without confabulations) in this control group only performed the pre- and post-measurements without treatment.

Healthy participants in this control group only performed the pre- and post-measurements without treatment.

Participants had to learn some brief material (words, faces, pictures, news), after which they were asked for an immediate and a delayed recall. After both recalls, participants were confronted with feedback about correct responses, non-responses and errors (i.e., confabulations and errors of attribution). This type of feedback worked on: 1) selective attention during the learning phase, training patients to focus on the relevant details of the stimuli; 2) monitoring processes during the retrieval phase, reinforcing the strategic search and training patients to inhibit traces that were irrelevant; and 3) memory control processes after the retrieval phase. The treatment consisted of 9 sessions and lasted for 3 weeks and the participants performed a baseline before and after treatment.

The confabulations recorded were 1) guessed answers, 2) confusions in time and space, 3) a mixture of two or more stimuli presented, and 4) devised or bizarre responses.

Scores ranged from 0 (no confabulations) to unlimited number of them (because devised or bizarre responses were recorded) and consisted of the sum of all the confabulations produced during the 3 sessions.

Scores ranged from 0 (no correct answers) to 72 (12 stimuli remembered twice in each session: firstly, in a immediate recall after learning, and secondly, in a delayed recall after 10 minutes).

Scores ranged from 0 (no non-responses) to 72 (12 stimuli remembered twice in each session: firstly, in a immediate recall after learning, and secondly, in a delayed recall after 10 minutes).

After the recall of the material, patients were also asked to remember which modality corresponded to each recall (i.e., seen, heard or imagined), and who had presented the material during the learning session (i.e., the therapist or themselves).

Scores ranged from 0 (if all answers were non-responses) to unlimited number (depending on number of confabulations produced by patients).