Typically, due to their structured character, cognitive rehabilitation tasks are disconnected from the situations people with schizophrenia meet in real-life. Indeed, everyday situations can be open ended, multitasking and engage a wide range of executive processes that are critical for the management of goals: the definition and the organization of subtasks according to an optimal order, the monitoring of multiple subtasks in parallel, the suppression of distracters, the detection of errors and their correction, delayed intention, and so forth. Many of these executive capacities, which are necessary for behavioral adaptation in daily life, are, however, very rarely addressed in laboratory exercises, yet difficulties with these capacities have been well documented in people with schizophrenia. For instance, Kessler et al. [8] identified deficits in specific executive functions (cognitive flexibility, inhibition of previously performed actions and control over a plan of actions) to explain everyday action disturbances in people with schizophrenia. Impact of executive impairments that is in attentional flexibility (shifting between self-generated goals and task-related stimuli), resistance to interference, and action sequencing abilities, was emphasized for particular activities of daily life (ADL) assessed in a real or computerized setting, such as choosing a menu, grocery shopping, and cooking a meal [9–11]. Surprisingly, in spite of their prevalence, goal-directed behavior difficulties meet by people with schizophrenia in their ADLs have not yet been the subject of cognitive rehabilitation programs, which in addition could orient the application of learned cognitive strategies in a real setting.

One technique from the field of rehabilitation for neurologically impaired people, Goal Management Training (GMT; [12, 13]), provides a direct focus on goal-directed behavior difficulties in daily life. GMT is based on Duncan's theory of “goal neglect” [14, 15], which addresses dysexecutive self-regulatory deficits (i.e., failure to carry out intentional and apparently well-memorized actions). The theory of goal neglect maintains that any activity requires a list of goals that are used to impose coherence and structure to behavior and to create a plan of actions, allowing one to achieve goals. During task execution, actual situation and stated goal are compared, and appropriate actions are then selected and activated in order to reduce discrepancy. Inhibition of actions not contributing to the achievement of a goal and selection of new actions when anterior actions fail to attain the goal (flexibility ability) are two other important aspects of the theory. According to Duncan [14], much of the disorganized behavior in people with frontal cortex dysfunction can be attributed to difficulties in elaborating and implementing such lists of goals.

The GMT procedure encompasses training 5 steps, each emphasizing one important aspect of goal-directed behavior: (1) stop: orient awareness toward the actual state of the situation; (2) define: the goal of the task; (3) list: the task into sub-steps; (4) learn: the steps; (5) check: if the result of an action corresponds to the stated goal. In the case of a discrepancy, the 5 steps are repeated. Therefore, the strategy consists of taking pauses during a current task to “stop and think”, selecting and maintaining the goal in memory while the task is executed, subdividing the goal in more simple and controllable subgoals, setting priorities for their execution, and using mental imagery or check lists. Exercises are administered in order to train people to use the strategy as a basis to evaluate their proper performance in ADL, implement new actions, and continuously monitor their success. GMT is a top-down approach that focuses on training processes that can be applied across domains and that encompasses different factors such as attention, problem definition, problem-solving, encoding and retrieval strategies, and monitoring. Thus, GMT aims to promote generalization of the training on activities not specifically addressed in a given intervention. The effectiveness of GMT has been demonstrated on several occasions such as in people with neurological lesions and in healthy elderly adults [13, 16–18]. In that people with schizophrenia also manifest goal-directed behavior difficulties in their daily life, GMT could prove beneficial for this clinical population as well.

Thus, the aim of this study was to assess the efficacy of Goal Management Training (with some modifications from the original GMT described in Levine et al. [16]) in a person with schizophrenia. A single-case methodology for the cognitive rehabilitation was used for several reasons. First, since executive processes underlying goal-directed behavior can be differently impaired from one person to another, and the consequences of the deterioration in ADLs can occur in various ways according to the person's interaction with his/her environment. Thus, a personalization of the intervention is essential in order to target the person's daily-life needs in relation with his/her cognitive profile (for an example, see [19]). In the present study, GMT was specifically chosen for a person who presented a cognitive and functional profile adapted to this type of intervention: he had executive problems in several everyday domestic tasks, which essentially consisted of difficulties in organization when faced with new and multitasking situations and in attaining the final goal; while at the same time, he presented efficient intellectual functioning, intact working memory processes (storage and processing load), and sufficient attentional capacities to cope with information. Finally, a single-case study was also favored because this methodology allows a detailed investigation of the person's (cognitive and functional) profile in order to isolate the specific components involved in the training.

The assessment consisted of the PROFINTEG instrument (general questionnaire) and a questionnaire about the use and the repercussions of the GMT strategy in G. O.'s daily life. They were administered to both G. O. and his reference caregiver in the sheltered house.

Regarding the postrehabilitation versus follow-up comparison of quantitative indices on the PROFINTEG questionnaire (see Table 1), the number of difficult activities remained stable (from 11 to 10), and the difficulty score decreased by 8 points (from 30 to 22). The decrease of difficulties concerned initiation and omissions. According to G. O.'s responses to the questionnaire, he prepared a meal without errors and help, and he varied the recipes that he prepared. G. O. carried out housework with less difficulty but he still doubted how well it was done. He explained that he did not mix different colored clothes when washing. The shopping list and shopping were still done together with fellow housemates, but he was able to do some shopping alone. He managed his appointments himself; however, financial and administrative management was still dealt with by the administrator. Finally, difficulties in concentrating during reading activity were still present. Furthermore, G. O. felt more conscious of his everyday functioning profile and more self-confident in the realization of everyday activities. He did not remember all of the 5 GMT steps but explained that he paid more attention to his environment and verified his actions (but not systematically).

Cognitive assessment showed that GMT had a beneficial effect on two planning tasks and an inhibition task. On the Tower of London (TOL), a task that assesses the capacity to analyze and elaborate possible solutions to a new problem, a decrease of the subsequent execution time, and the number of the moves to reach the solution were observed. These results could be explained by a more fully formed plan, which resulted in improved accuracy. On the 6 Elements Test, G. O. had to attempt to do 6 simple subtasks within a limited time period, while at the same time respecting a number of rules. The goal was to apply an effective strategy, that is, engage in time management throughout the 6 subtasks and initiate task switching in order to complete the first items. After the rehabilitation, G. O. was able to organize the rules in order to perform all the subtasks efficiently. Finally, on the Hayling test, which assesses the ability to inhibit a strongly cued automatic response and to find an unrelated one, the number of errors greatly decreased. Before the rehabilitation, the observed deficit of response suppression could be due to the persisting use of routine semantic schema and a decrease in strategy use [33]. This observation could also be related to G. O.'s poor performance on the TOL at prerehabilitation where he showed an inability to inhibit an inappropriate move that was very strongly triggered by the context [34]. Thus, these results suggest that GMT improved G. O.'s ability to inhibit schemas induced by the context. Finally, the score on the errand test was impaired after rehabilitation. This may be due to the fact that, for this task, it was too difficult for G. O. to keep all the instructions in mind (i.e., make a plan and identify alternatives approaches).

Regarding the everyday paper-and-pencil tasks, proofreading and grouping, improved performance on these tasks were also associated with GMT. Indeed, G. O. read the instructions and performed the tasks more slowly and made fewer omissions. These results could indicate that GMT increased G. O.'s care and attention to the tasks, which in turn reduced errors. These results replicate those observed in Levine et al. [16], furthermore suggesting the specificity of the effects of GMT on these tasks. The beneficial effect was observed on the task (proofreading) that was incorporated into training, as well as on the task (grouping) that was not specifically addressed by the intervention. This result indicates that there was a generalization of the training effects to the laboratory tasks. On the contrary, no change was observed on the room layout task that was incorporated into training. Here, G. O. expressed difficulties in listing the substeps in order to formulate a plan.

Beneficial effects of GMT were also demonstrated qualitatively and quantitatively on two ADL assessed in a real setting: meal preparation and meeting preparation. A considerable decrease of errors (omission, addition, and inversionsubstitution) was observed, and G. O. was less dependent on the examiners. Moreover, the analysis of errors provided evidence of the different mechanisms underlying the GMT effects. First, the reduction of context neglect errors and the more frequent consultation of the instructions and rules indicated that G. O. took the contextual information more into account. This could indicate that there was an improvement in his ability to analyze and deal with the environment, the first step of problem resolution [23]. This step has been compensated by GM strategy, which taught G. O. to direct his attention toward pertinent information and inhibit inappropriate responses. A less redundant consultation of the recipe in the meal preparation task and of the guest list in the meeting preparation task might also suggest that G. O. was able to better structure his interaction with information. Additionally, this context information processing enabled G. O. to revise an initial plan in the face of the external contingencies. For instance, in the meeting preparation task, G. O. removed the name card and material of the guest who withdrew from the meeting. Second, G. O. demonstrated that he was able to monitor ongoing actions, for instance in the meal preparation task where he controlled the cooking time of cake. Third, G. O. often used the checking step of GM strategy, allowing him to reduce the number of errors. This was particularly clear during the meeting preparation task, where he checked and monitored his actions very actively by reading the rules and by orally reciting them. Finally, he spontaneously made commentaries about actions that he was doing or that he had to do, which could be a sign of organization and verbal self-regulation [35, 36].

On the whole, these findings suggest specific effects of GMT on targeted processes: context information processing, ongoing response monitoring, checking, and verbal self-regulation. It is unlikely that these effects are due to mere practice on the tasks without having learned a specific strategy [13]. Moreover, on the control task that involved divided attention, the pattern of performance remained impaired indicating no general effect of GMT. The absence of improvement on different cognitive functions (flexibility, and attentional functions) that were impaired at prerehabilitation also demonstrates the specificity of GMT in our study.

Generalization of GMT effects to other real-life contexts was manifested as improvement was observed on the meeting preparation task, a task that was not targeted during the cognitive rehabilitation program. Several issues can account for this effect. GMT is conceived as a top-down strategy, which can engage different behaviors and be applied to various contexts that need plan formulation. Moreover, generalization of learnt strategies was built into the intervention. Indeed, G. O. learnt how to apply GMT in contexts other than the target activity (i.e., meal preparation), such as domestic chores and different situations as illustrated in homework exercises. Finally, psychoeducation given before the training was very important in order for G. O. to become aware of the cognitive mechanisms involved in various ADL.

G. O.'s self-esteem also improved after the intervention. For instance, before the rehabilitation, G. O. expressed lack of self-confidence about the completion of ADL. The implementation of the GM strategy in his daily-life enabled him to directly improve his self-confidence. On the contrary, results regarding clinical symptoms (the latter were not severe to begin with) did not show any significant improvement after the rehabilitation.

One key factor in the success of GMT was no doubt related to G. O.'s high level of motivation throughout the rehabilitation program (based on comments made by G. O. and the fact that he attended all the sessions). Motivation and rehabilitation engagement were in part promoted through the personalization of the rehabilitation goal and material, training of ADL, and verbal encouragement from the therapist. Indeed, according to Medalia et al. [5, 37], intrinsic motivation (desire to engage in an activity because it is inherently interesting and engaging) is an essential factor to consider as in a learning environment it could be associated with greater learning, higher self-esteem and a sense of well-being, and greater engagement.

Some negative results following the application of GMT on the ADL deserve further comment. First, the beneficial effects of the GMT did not generalize to the washing activity. This might be related to the fact that this task was familiar to G. O. Indeed, he washed clothes once a week and he was used to doing it without sorting out according to temperature and color of clothes. This routine action relied thus on schema that specified in detail how the behavior should be carried out. In order to improve or change the behavior, G. O. had to reject the existing schema and create a new schema [23]. Limited impact of GMT on the washing task could be explained by G. O.'s difficulty in inhibiting the routine schema. It might be more difficult to spontaneously transfer training effects to routine actions, in contrast to new and nonroutine actions (such as the meeting preparation task), as new skills can be more easily and quickly incorporated [38]. Second, goal definition and splitting up of a task in substeps remained difficult steps to carry out by himself. In particular, G. O. had difficulty in mentally simulating the real execution of an action plan, which could be related to an observed defective ability in action sequencing, that is, detecting boundaries in large action units, such as macrosteps, and setting priorities among the events with regard to the stated goal. These difficulties could indicate a disturbance in causal connections between the component actions to represent a plan as a coherent and structured sequence of goal-related events [39, 40]. Consequently, the increased amount of fragmentized actions could have overloaded G. O.'s working memory capacity during the execution and monitoring of actions in ADL. GMT aimed to remediate these difficulties by teaching G. O. to define the main goal and to sequence the main goal into substeps by selecting the most appropriate steps for achieving the goal. G. O. also learnt to estimate and manage the time for the articulation of action plans. Moreover, when the steps of an activity were defined on a sheet of paper, G. O. could be guided by their automatic execution and monitoring (by following and crossing out each completed step on a sheet), decreasing the demands on working memory, and the intervention of voluntary control. Third, purposeless actions increased on the meal preparation task indicating that hyperactivity was still present in G. O. This could be related to his high level of anxiety that did not change after the rehabilitation and was predominant in his symptomatology. Finally, the caregiver's responses on the PROFINTEG questionnaire regarding G. O.'s daily functioning did not indicate a substantial decrease in difficulties, unlike G. O.'s responses. It should be noted that it was difficult for the caregiver to perceive everyday changes, as he only had brief contacts with G. O. mainly due to time constraints.

Regarding the two-year followup, the durability of the beneficial effects was largely evident in that the beneficial effects of the cognitive rehabilitation program were still present. Indeed, G. O. still had fewer difficulties in the realization of everyday activities at followup. Additionally, G. O. expressed being more self-confident and autonomous—crucial goals of a rehabilitation program. However, it is important to note that a few rehabilitation sessions aimed at refreshing acquisition would have been necessary in order to favor a better internalization of the 5 GMT steps.

Several limitations can be mentioned. First, blind assessments were not carried out. However, this was not feasible as G. O.'s assessments were carried out as part of a cognitive rehabilitation program. We considered filming G. O. while performing the various tasks, but he did not approve of this. Nevertheless, all assessments were always conducted by two judges. Second, there was a lack of stability of performances on the proofreading and room layout tasks at pre- and postrehabilitation.