Introduction

The properness of our past actions is judged according to counterfactual possibilities; that is, what could have been if we had behaved differently. The ability to simulate alternatives to factual events is called counterfactual thinking (CFT), and it is a signature of humans’ cognition [1]. Conditional propositions are typically used to think counterfactually, often converting unusual behaviors into more “normal” antecedents [1]: “If I had taken the usual way, I would not have missed the flight”. The closer the outcome is to the expected (missed) goal, the higher the probability to elicit CFT [1].

But why do we think counterfactually? The acknowledgement that a better (upward) outcome could have been achieved elicits regret, guilty, and self-blame; the greater the salience of the missed achievement, the worse the judgment [2, 3]. Nevertheless, we might learn an important lesson for the next time. Indeed, the most prominent function of upward CFT is to provide useful insight about possible alternative ways to achieve missed goals, influencing future behaviors [4, 5]; obviously, as long as we recognize that what we did (or did not) was the causal antecedent of our failure. On the contrary, downward counterfactuals denote worse possible scenarios, prompting a sense of relief, thus, having a beneficial affective-regulation purpose [6].

Consequently, impaired CFT might have deleterious effects on the ability to modulate both behaviors and affections. Lack of CFT could impact behavioral regulation preventing individuals’ experiential learning and determining underachievement and/or social dysfunction; conversely, excessive CFT may lead to unnecessary worries, anxiety, and dysphoria [4, 5]. Thus, CFT difficulties might be somehow related to all those clinical conditions characterized by intense psychological distress, and emotional or behavioral dysfunctions. In support of this hypothesis, meta-analyses showed that there is a significant association between CFT hyperactivation and post-traumatic stress disorder (PTSD) [7] and between upwards counterfactuals generation and depression [8]. Furthermore, altered CFT was observed in other psychiatric conditions, such as schizophrenia [9, 10] and the obsessive compulsive disorder (OCD) [11]. The causal direction of these associations and the role of possible moderating factors are still under debate [4, 7, 8]. On the other hand, the literature suggests quite strongly that CFT difficulties might be related to fronto-executive vulnerabilities, at least concerning depression and schizophrenia [4, 12, 13]. Indeed, CFT impairments have been observed also in neurological conditions characterized by a fronto executive dysfunction, such as Parkinson’s Disease (PD) [14], Huntington’s Disease (HD) [15], and frontal lobe damages [16].

To sum up, CFT difficulties may be detected in a wide range of psychiatric and neurological conditions, with possible negative effects on people’s affective and behavioral regulation. Accordingly, the acknowledgement of the clinical populations possibly involved might be essential for optimal assistance and proper clinical management. For instance, a more refined understanding of the possible role of counterfactuals in maintaining PTSD and depressive symptoms might be crucial for tailoring efficacious interventions [7], which could address self-compassion and aim at minimizing the contemplation of negative outcomes [8]. Furthermore, in those conditions characterized by behavioral dysfunctions, such as PD [17], HD [18], schizophrenia [19], and cerebral frontal damages [20] it might be extremely valuable to identify the possible contribution of CFT impairments on behavioral regulation and experiential learning, to define possible compensating strategies. To this purpose, a comprehensive acknowledgement of those conditions possibly affected by CFT deficits and the understanding of how these difficulties might be related to the clinical picture is recommended.

As illustrated, psychiatric and neurological conditions have been identified as the best candidates for possible CFT impairments [4, 5]. Systematic reviews and meta-analyses have been provided for depression [8] and PTSD [7]; however, no attempt was made to collect and summarize what we know so far about possible CFT difficulties across several different conditions in adulthood. Indeed, such investigation might provide useful guidance for researchers interested in the topic, to be used as a reference point for future works. For this reason, a scoping review was made to provide a narrative overview about the state of art in the field, identify the methodologies adopted, the variety of psychiatric/neurological conditions already considered, and the relevant issues to be addressed by future investigations.

Methods

The review protocol was developed according to PRISMA extension for Scoping Reviews (PRISMA–ScR; PRSIMA-ScR Checklist is provided in S1 Table) [21]. The final protocol may be available on request. Original articles published in peer-reviewed journals at any time were included if they were written in English and if they quantitively and/or qualitatively measured CFT in adult participants (i.e., > 18 years old) affected by psychiatric or neurological diseases. Records for which the full text was not available were not included, since the impossibility to get detailed information.

Records were identified by searching the most relevant electronic databases in the field: PsycInfo, PubMed, Scopus, and Web of Science. The last search was run on 23^rd January 2020. Specifically, we searched the term “counterfactual thinking” or “reasoning” or “thought” in combination with the following terms: “disease”, “illness” “clinical population”, “clinical sample”, “impairment”, “deficit”, and “patient”. The detailed path used in each database is reported in S1 Appendix.

One reviewer performed the search and made a preliminary eligibility check based on titles and abstracts after removing duplicates. Then, the full text of all the other papers was analyzed in detail, and eligibility was discussed with a second reviewer. If needed, disagreements were resolved by consensus and discussion. Furthermore, the reference lists of the full texts considered were screened to identify additional pertinent articles.

Data were collected by one reviewer according to an extraction sheet that was previously developed with another reviewer. Specifically, the following information was collected for each work: i) characteristics of the sample including disease type and sample size, ii) the tasks administrated and iii) the main results. All data sought were found in the original articles, no additional research was needed.

First, we will provide a brief description of the tasks used and the clinical conditions considered in the reviewed studies that may rapidly inform researchers who want to extend the previous evidence in this field. Then, we will report the main findings grouping the records by the condition considered, that is psychiatric or neurological. Moreover, in each of these two sections studies will be further grouped according to the specificity of the diseases. More specifically, distinguishing between studies on schizophrenia, depression, post-traumatic stress disorder (PTSD), anxiety and obsessive-compulsive disorders, autism spectrum disorders, and personality traits; concerning the neurological section, involving patients with cerebral lesions, neurological syndromes, or neurodegenerative disorders.

Results

The selection process is reported in Fig 1. The search performed on PsycInfo, PubMed, Scopus, and Web of Science provided 252 citations. Additionally, 15 studies were detected by manually inspecting the reference lists of the full text papers analyzed. After the removal of duplicates 150 citations remained. All of them were screened for eligibility by title and abstract. A total of 110 records were excluded since eligibility criteria were clearly violated (e.g., studies on children, or different clinical conditions, or not written in English). The full text of 40 works was analyzed in detail. Six studies were excluded since the full text was not available and five were excluded since they did not measure CFT. A total of 29 studies were identified for inclusion in the review. Considering the financial support received, fourteen studies (48%) were founded by public agencies, two (7%) were supported by private foundations, five (17%) received miscellaneous support by both private and public agencies, one (4%) received no support and seven studies (24%) did not specify information about financial support.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The flow diagram represents the studies selection process.

* Six studies were excluded since the full text was not available and five were excluded since they did not measure CFT.

https://doi.org/10.1371/journal.pone.0246388.g001

The clinical condition considered in each study, sample size, task administered, and main results are reported in Table 1. Twelve studies (41%) probed the generation of counterfactual thoughts related to a negative real-life [14–16, 22–27] or a fictional event [9, 23, 25, 28, 29]. Ten studies (34%) administered the Counterfactual Inference Test (CIT) [22] to investigate the ability to use CFT to make causal inferences in fictional social scenarios [12, 14–16, 22, 24, 28–31]. Six works (21%) used the Counterfactual Thinking for Negative Events Scale [32, 33] or similar questionnaires [34–37] investigating the content (e.g., upward/downward) and the affective aspects of CFT. Four studies (14%) investigated CFT indirectly through a counterfactual gambling task [38] measuring the experience of regret in those circumstances in which participants win less (or loss more) than how much they could have won (or lost) making a different gambling choice [11, 12, 31, 39]. Indeed, the experience of regret is a marker of an efficient CFT since it grounds on the ability to figure out how things could have gone differently if a dissimilar choice had been made. Likewise, two studies (7%) [15, 24] investigated the influence of anticipated counterfactual regret on behavior [40] asking participants to say how they would have behaved in regret-inducing imaginary circumstances. Individuals who implement an effective CFT are expected to behave anticipating future regret. Finally, two studies (7%) probed the “causal order effect” [41], meaning the tendency to choose the first event of a sequence as the most decisive for the final negative outcome [9, 28]. Finally, two studies (7%) investigated the influence of CFT on the automatic activation of behavioral intentions using the sequential priming paradigm [10, 30].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of studies included in the review.

https://doi.org/10.1371/journal.pone.0246388.t001

Overall, 21 out of 29 studies (72%) focused on psychiatric diseases whereas only eight studies (28%) considered neurological conditions. More specifically, six works focused on schizophrenia [9, 10, 12, 22, 28, 30], five probed CFT in the presence of depressive symptoms [13, 23, 26, 27, 32] and five investigated the relationship between CFT and post-traumatic distress [25, 32–34, 42]. Three studies addressed the relationship between CFT and anxiety traits [32, 35–37]. One focused on the obsessive-compulsive disorder [11], one on autism spectrum disorders [31], and one probed CFT in miscellaneous psychiatric conditions, including schizoid and schizotypal traits [32]. Finally, one work probed CFT in participants with different levels of psychopathy [43]. As concern neurological diseases, solitary studies focused on prefrontal cortex (PFC) [16] orbitofrontal cortex (OFC) [38], and hippocampal damages [29], Parkinson’s [14] and Huntington’s [15] diseases, relapsing-remitting multiple sclerosis [39], Tourette syndrome [24], and chronic pain [44].

Discussion

A scoping review was conducted to provide a narrative overview about previous evidence on CFT in adults with psychiatric or neurological diseases, identifying the methodologies adopted, the variety of psychiatric/neurological conditions already considered, and the relevant issues to be addressed by future investigations. Overall, twenty-nine studies have been reviewed. The generation of counterfactual thoughts related to a negative real-life [14–16, 22–27] or a fictional event [9, 23, 25, 28, 29] and the CIT [12, 14–16, 22, 24, 28–31] were the most popular tasks adopted, followed by the Counterfactual Thinking for Negative Events Scale [32, 33] or similar questionnaires [34–37] and the counterfactual gambling task [11, 12, 31, 39]. Only a minority of studies investigated the influence of anticipated counterfactual regret on behavior [15, 24], the “causal order effect” [9, 28], and the influence of CFT on the automatic activation of behavioral intentions using the sequential priming paradigm [10, 30].

Overall, a wide variety of clinical conditions were investigated; however, most studies focused on psychiatric diseases and specifically schizophrenia, depression, and PTSD. Concerning patients with schizophrenia previous findings point to the possible role of reduced CFT [9, 22, 28], low regret sensitivity [12], and qualitatively different causal attributions in favoring suboptimal behaviors [9, 10, 22, 28]. Considering depression, results are a bit more confusing. On one side, it emerged that depressed individuals might be less inclined to reason counterfactually [13, 23] and that they might not take advantage of CFT either to prepare for future similar situations [23] or as an affective regulator [27]. However, other studies reported that depressed individuals generate more pessimistic and fewer optimistic counterfactuals [32] and that they focus significantly more on self-related controllable aspects of events; possibly, empowering their low perception of personal control but, also, incrementing the perceived distress when depression is severe [26]. Thus, these findings only partially agree with the results of Broomhall and colleagues’ meta-analysis (2017), showing that a higher predisposition to upward counterfactual thinking and regret “may serve as precursor to depression”. Nevertheless, their systematic search enabled to identify more studies than those discussed in the present work, for instance including research in which the evaluation of depressive symptoms was a secondary outcome (e.g., in terminally ill patients or early motherhood). However, methodological factors (e.g., sample type, instruments used) likely affect the observed relationship between CFT and depression, as they reported. Therefore, the inclusion of additional studies might explain dissimilar conclusions. Furthermore, they considered only upward counterfactuals, whereas no distinction was made in the present work. On the other hand, concerning CFT in PTSD, our report is in line with a recent meta-analysis on the topic [7] suggesting that in PTSD a high frequency of intrusive and dysfunctional counterfactuals about the trauma might exacerbate post-traumatic distress [25, 33, 34, 42]. Thus, therapists should be acknowledged of the possible role of impaired CFT in sustaining post-traumatic symptoms. Finally, a minority of studies assessed CFT in other psychiatric conditions or in the presence of specific personality traits, reporting anomalous CFT generation associated with schizoid and schizotypal [32] and anxiety traits [35–37]; atypical regret sensitivity in AS/HFA [31] and a reduced tendency to modulate one’s choice behavior anticipating future regret in psychopathy [43] and OCD [11]. Furthermore, few and isolated studies considered CFT in neurological conditions, showing that patients have a reduced tendency to think counterfactually [14–16] and to experience regret for their choices [38, 39] when the diagnosis involves a fronto-executive dysfunction. Nevertheless, the effect of altered CFT on patients’ behaviors [15, 24, 38, 39] and the use of CFT to make causal inferences might be differentiated according to the specificity of the disease [14–16, 24, 29].

To sum up, CFT seems to be altered in several psychiatric and neurological diseases, even though with some specificities across the different clinical conditions, and it possibly affects patients’ affections and behaviors. Besides, what emerges clearly is that CFT deficits might be specifically associated with a fronto-executive dysfunction, likely related to damages of the PFC and OFC [12, 13, 16, 38, 39], in both psychiatric and neurological conditions. Indeed, the PFC and, more specifically, the OFC are known to be crucially involved in the executive control [48] and they underpin the simulation of counterfactual alternatives in healthy individuals [46, 47]. Additionally, the OFC supports the integration of emotion and cognition in relation to decision making and behavioral planning [49]. Thus, PFC and OFC damages and the presence of fronto-executive dysfunction might be considered crucial risk factors for impaired CFT.

Nevertheless, possible limitations related to our review process might be noted. As mentioned, we did not perform a systematic qualitative and quantitative investigation; thus, possible relevant studies might be omitted, and results have been reported in a narrative fashion. However, a rigorous methodology was followed according to PRISMA guidelines for scoping review; thus, in line with the standards posed by the international scientific community. Also, we included only studies written in English possibly limiting the coverage of our investigation. Nevertheless, English is adopted internationally by the scientific community; thus, we expected more relevant research to be published in this language. Moreover, we focused specifically on psychiatric and neurological diseases in adults, although CFT impairments might be detected also in other diseases, as well as in children. Therefore, future studies might consider reviewing the evidence about different clinical populations, including developmental conditions.

Concluding, the previous research on CFT in the clinical context deserves credit for highlighting the presence of possible CFT impairments in several psychiatric and neurological populations, also drawing quite strong evidence about the neural networks involved. However, studies sometimes lack methodological rigor. For instance, some works did not include a control sample of healthy participants [32, 33, 42–44], did not involve participants with clinically relevant symptomatology [23, 26, 27, 35–37, 43], or did not clearly distinguish between clinically significant and non-significant symptoms [33, 42]. This might be especially risky since counterfactual biases and suboptimal inferential processes have been reported also in the healthy population [50]. Furthermore, four studies included only female [35–37, 44] and one included only male [43] participants; however, most of the studies controlled for sex and/or other demographic variables such as age and education. Additionally, even though a wide variety of clinical conditions was investigated, most studies focused on psychiatric rather than neurological diseases and specifically, schizophrenia, depression, and PTSD. Therefore, future research might further explore the possible relationship between CFT and the other conditions in which CFT difficulties have been preliminary reported, such as anxiety, OCD, personality traits, autisms spectrum disorders, and psychopathy. Likewise, considering the paucity of evidence about CFT in neurodegenerative disorders, we might recommend further investigations also in this field. Specifically, we might note that given the crucial role of the fronto-executive functioning, the investigation of CFT should be extended to all those conditions characterized by fronto-executive impairments. For instance, Amyotrophic Lateral Sclerosis (ALS) might be of specific interest since 20% to 50% of ALS patients develop Fronto-Temporal Dementia (FTD) [51, 52], including a lack of executive control [53]. On the other hand, since quite a number of studies have been made on CFT in schizophrenia there might be room for a systematic qualitative and/or quantitative analysis (as done for depression and PTSD). Finally, one might speculate that as CFT plays a role in driving everyday behaviors, it might be crucial also when medical decisions are involved. Indeed, neurological and psychiatric patients likely face complex and thorny medical and therapeutic choices, possibly with relevant medical and legal implications. However, the possible effect of impaired CFT on decision making in the clinical context has never been investigated. Future studies might address this issue. Furthermore, future research should focus on the development of interventions aimed at compensating for patients’ possible CFT difficulties.

Supporting information