Overall cognitive assessment. Several tasks were used in order to obtain a general profile of cognitive functioning. The Mini Mental State Examination (Rovner and Folstein, 1987) was used as a basic screening for cognitive impairment. The capacity to sustain attention and divide attentional resources between tasks was assessed using the Telephone Search (TEA, Robertson et al., 1994). The ability to comprehend verbal instructions was measured using the Token Test (De Renzi and Faglioni, 1978). A logic memory task (Wechsler, 1987) was used to assess participants’ capacity to register and recall new verbal information. Similarly, the Rey-Osterrieth Figure (Stern et al., 1994) was used to assess visual memory as well as visuo-spatial abilities. Finally, the Frontal Assessment Battery (Dubois et al., 2000) was employed as a screening for executive abilities.

Cognitive control assessment. A set of neuropsychological tasks was used to obtain a profile of several cognitive control abilities:

Emotional functioning assessment. In order to assess the presence of symptomatology, the Hospital Anxiety and Depression scale (HADS; Zigmond and Snaith, 1983), a self-report questionnaire, was employed. The HADS has demonstrated to be a sensitive tool to assess depression and anxiety symptoms in acquired brain injury population (Dawkins et al., 2006). To assess the use of reappraisal in daily life, the Emotion Regulation Questionnaire (Gross and John, 2003) was administered.

This task is adapted from several studies on reappraisal ability, and is described in detail in Salas et al. (2013). 10 pictures¹ were selected from the International Affective Picture System (IAPS; Lang et al., 2008), depicting negative events of different sorts. These pictures were selected to cover the wide range of possible negative scenarios that people commonly face (death, natural disasters, accidents, illness, violence, etc.). The pictures were displayed in a 14′ computer screen.

At the beginning of the task participants were trained to generate reappraisals, using three practice IAPS pictures (not part of the set of 10 test pictures). The task was introduced as follows: “Sometimes people try to feel better by looking on the bright side of things. You will watch pictures of negative events and will be asked to think aloud about the positive side of these situations. Try to be fast and say as many positive sides you can think of.”

In order to avoid memory bias for the neurological group, on the computer screen, above each picture, the task instruction was summarized: “Think aloud about the positive side of this situation. Try to be quick.” Participants were informed that their answers would be timed, and recorded verbatim. They were also informed that the aim of the task was to produce as many positive reinterpretations as possible. If they were not able to generate a correct reappraisal for the first picture, several reappraisal examples were offered [e.g., Car Crash (9903): “when looking at this picture some people say that help is on the way” or “is not as bad as it looks”]. The same procedure was followed with the second and third practice pictures. Both the neurological group and the non-brain injured group were able to offer adequate reappraisals by the end of the three-picture practice phase.

The average number of seconds taken to generate a first reappraisal did not differ between the LH and RH groups. However, significant differences were found between patients with brain injury in general (LH + RH), and HC. A Kruskal–Wallis non-parametric test was used to compare the time taken by each group. It was observed that the number of seconds was significantly different between groups [H(2) = 10.77, p = 0.002; mean ranking: LH = 21.71, RH = 18.25, HC = 9.79]. According to planned comparisons, it was found that this difference was only significant, and had a large effect size, between the HC group and the brain injury group (LH + RH; U = 32, Z = -3.19, p < 0.001, r = 0.59), but not significant, and with a negligible effect size, between LH and RH groups (U = 25, Z = -0.24, p = 0.43, r = -0.04). In conclusion, our findings do not support the LH Reappraisal Hypothesis, for patients with LH and RH damage were equally slowed in generating reappraisals.

The average number of reappraisals generated for each picture did not differ between groups, suggesting that individuals with LH and RH lesions, and healthy controls, are equally productive when time is not considered. A Kruskal–Wallis non-parametric test was used to compare the total reappraisals generated by each group. It was observed that the number of reappraisals was not significantly different between groups [H(2) = 3.51, p = 0.175; mean ranking: LH = 14.07, RH = 10.88, HC = 17.82]. A detailed description of the groups’ performance can be found in Table 3. A graphic summary of the results can also be found in Figure 1.

From the four cognitive control processes considered, only two (inhibition and verbal fluency) were significantly associated with reappraisal difficulty. The final model tested had reappraisal difficulty as a dependent variable and working memory, abstraction, verbal fluency, and inhibition as predictors. Multicollinearity, homoscedasticity, and independency of errors assumptions were met. The model was significant [F(4,24) = 7.70, p < 0.001] and explained a 49% of the variance. The final model is presented in Table 4. From the four predictors, inhibition and verbal fluency were the only two significantly associated to reappraisal difficulty. By looking at the standard coefficient it is possible to conclude that Inhibition showed the strongest predictive value. In relation to the other two independent variables, abstraction was marginally related and working memory showed no relationship.

The same model was tested for reappraisal productivity. Contrary to the previous model, this one was not significant [adjusted R² = 0.022, F(4,24) = 1.16, p = 0.35].

Our focus in this study was reappraisal generation, or the capacity to produce positive reinterpretations of negative events. This is an important point to clarify, because most behavioral and neuroimaging studies to date have focused on reappraisal ability, or the capacity to modulate emotion through the use of reappraisal. Such emphasis on ability is sensible, considering that people with healthy brains probably have an intact capacity to generate reappraisals. However, a robust set of evidence on the neuropsychological consequences of brain injury suggests that the generation and manipulation of thoughts can be compromised by lesions to diverse brain areas (Goldstein and Scherer, 1941; Luria, 1966; McGrath, 1991; Gomez Beldarrain et al., 2005). If reappraisal relies on the flexible use of thinking (Ochsner and Gross, 2004), then it is likely that people with brain injury may struggle using such strategy to modulate how they feel.

Data from this study suggest that lesions to the LH, which has been long related to language functions (Frost et al., 1999; Corballis et al., 2012), and verbally mediated cognitive control (Stephan et al., 2003; Gruber and Goschke, 2004; Henry and Crawford, 2004), do not impair reappraisal generation more than lesions to the RH. This is interesting, since reappraisal has been mostly considered as a language- mediated ER strategy (Gross and Thompson, 2007; Mcrae et al., 2012a). A possible explanation for this negative finding may be related to the fact that LH and RH groups presented a similar profile of cognitive impairment (see Table 2). It is also possible that individuals with RH lesions presented low scores on reappraisal difficulty as a consequence of impairment of other non-language mechanisms, for instance, attention. According to our data, both groups are impaired on their ability to sustain attention, a deficit that is commonly expressed in everyday life as distractibility (Leclercq et al., 2002). If reappraisal involves the manipulation of attentional focus (Ochsner et al., 2002), it is possible that attentional deficits may have an impact on reappraisal generation. However, this explanation remains speculative and requires scientific exploration.

An alternative explanation for this negative finding may well be related to the small sample size of the study. Even though there is a small difference between LH and RH groups, suggesting that people with LH lesions are slower than RH individuals generating reappraisals, this does not appear significant. In addition, this negative finding may also be interpreted in view of the unbalanced distribution -inside each neurological group- between anterior and posterior lesions. In fact, the LH group presented less individuals with prefrontal damage (anterior = 3, posterior = 5) compared to the RH group (anterior = 7, posterior = 1), perhaps compromising comprehensive (posterior), and not executive (anterior) aspects of language. In the future, new studies should consider this variable and generate experimental designs that include individuals with lesions to anterior and posterior areas of each hemisphere. Considering the existing neuroimaging literature on reappraisal (see Introduction), it is possible that damage to the anterior portions of the LH will have a more marked impact on the capacity to generate reappraisals than damage to posterior areas of the left convexity.

The available literature on cognitive control and reappraisal is scarce and focuses exclusively on reappraisal ability. So far, it has been suggested that reappraisal ability is positively related to working memory (Schmeichel et al., 2008; Mcrae et al., 2012b) and marginally related to abstraction (Mcrae et al., 2012b). In relation to verbal ability (e.g., to generate alternative interpretations), only one study has found indirect evidence that verbal fluency is related to reappraisal ability, however, this on a suppression task (Gyurak et al., 2009). Interestingly, no associations have been reported for inhibition, a key theoretical aspects of reappraisal (e.g., to detach from the negative emotional experience).

This study suggests that an inhibitory control task (Drewe, 1975), where participants must withhold a response that was given to the same stimulus before (e.g., tap once when the examiner taps twice), can significantly predict the amount of time that it takes to generate a first reappraisal. In other words, if an individual decreases in one point his score in the inhibition task (Dubois et al., 2000), his response time to generate a first reappraisal will increase by 10.27 s. This finding is the first to support the view that inhibition is a key ability in decreasing the salience of automatic negative appraisals (Mcrae et al., 2012b), for when inhibition is impaired, reappraisal generation requires considerably more time. There is one case study that has reported this relationship in detail (Salas et al., 2013), describing how inhibition impairment after a left fronto-parietal lesion produced a remarkable difficulty to spontaneously generate reappraisals. In addition, this finding is also consistent with a large literature associating damage to the left and right PFC with deficits withholding prepotent responses (Aron et al., 2003, 2004; Swick et al., 2008). Finally, it is important to note that this data is also consistent with models of executive function that propose behavioral inhibition as a requisite to any goal directed behavior (Barkley, 1997, 2001).

The results obtained in this experiment also suggest that performance on a verbal fluency task (Delis et al., 2001), where participants are expected to generate as many words that begin with a letter in one minute, also predict the amount of time taken to generate a first reappraisal (albeit to a less extent than inhibition). Verbal fluency is an interesting measure to consider in the context of reappraisal, both because deficits in verbal fluency are a common feature of prefrontal cortex damage (Henry and Crawford, 2004; Robinson et al., 2012), and also because verbal fluency presents high associations with verbal ability (Miller, 1984; Crawford et al., 1992, 1993; Sollberger et al., 2010), a core component of reappraisal (Mcrae et al., 2012b). In addition, verbal fluency also appears to recruit other processes that are key to reappraisal, such as the retrieval and recall of information, self- monitoring and inhibition (Perret, 1974; Henry and Crawford, 2004).

Based on these findings, and considering data from previous studies on reappraisal ability, a two-stage reappraisal model may be proposed (see Figure 2). In a first stage (reappraisal generation) inhibition is required to disengage from the automatic negative meaning. If inhibition is successful, alternative interpretations, or new meanings, can be generated -this moderated by verbal fluency. In addition, and considering previous data (Schmeichel et al., 2008; Mcrae et al., 2012b), it is also suggested that, during a second phase (reappraisal maintenance), working memory ability has a role keeping in mind the recently generated new interpretation, thus “shielding” it from the initial meaning that still remains in the focus of attention (Kanske et al., 2010; Gross, 2013). This model is consistent with the implementation-maintenance model of reappraisal, IMMO (Kalisch, 2009), which suggests that the process of reappraisal has early and late components. According to this model, early components refer to operations needed for choosing and implementing an initial reappraisal strategy, while late components are required to maintain a reappraisal in working memory and monitor its success. Taken together, data from this lesion study suggests that inhibition and verbal fluency, two neuropsychological abilities, are critical in implementing an initial reappraisal strategy during the early phase of the process. Working memory, on the contrary, seems not to be relevant in generating (implementing) a reappraisal during the early moments of the process.

This study offers insights into how brain damage may compromise ER. It suggests that brain injury, without regard to lesion laterality, impairs the capacity to quickly generate positive reinterpretations of negative events. However, when longer periods of time are provided, patients are able to produce reappraisals as well as controls. These findings are relevant to neuropsychological rehabilitation for two reasons.

First, they offer supporting evidence to the idea that brain-injured patients in general are vulnerable to experience emotion dysregulation (Salas, 2012), particularly when confronted to emotional situations where they need to react quickly. This is consistent with studies describing that focal and non-focal brain damage compromises the capacity to rapidly react to environmental demands (Goldstein, 1995; Gerritsen et al., 2003; Winkens et al., 2006; Mathias and Wheaton, 2007), and that negative emotional events can compromise cognitive abilities (e.g., Demanet et al., 2011).

Second, they suggest that a psychological capacity such as reappraisal can be externally modulated, this, by manipulating environmental demands (time). In other words, people with acquired brain injury are slower than controls, but if enough time is offered, they are equally able to generate reappraisals, at least in the case of relatively high-functioning individuals such as those studied in the present research. This is in line with evidence proposing that neuropsychological impairments are not stable, but can be modulated by physical or interpersonal context (Bowen et al., 2010, for a review). For example, it has been reported that the use of prompts facilitates the process of disengagement from negative visual material, increasing dramatically the capacity to generate positive reinterpretations (Salas et al., 2013). This evidence appears to support the theoretical proposition that extrinsic forms of ER, such as affective and cognitive engagement (Niven et al., 2009), can be used to compensate for intrinsic ER difficulties (reappraisal; Freed, 2002; Salas, 2012).