We support the approach of Lane et al. in focusing on cognitive processes in understanding psychopathology and how to treat it. However, we believe a broader range of processes is necessary to address in particular cases. In our papers (McIntosh et al. Reference McIntosh, Sedek, Fojas, Brzezicka-Rotkiewicz, Kofta, Engle, Sedek, von Hecker and McIntosh2005; Sedek & von Hecker Reference Sedek and von Hecker2004; Sedek et al. Reference Sedek, Brzezicka, von Hecker, Gruszka, Matthews and Szymura2010; von Hecker & Sedek Reference von Hecker and Sedek1999; von Hecker et al. Reference von Hecker, Sedek and Brzezicka2013) and edited monographs (Engle et al. Reference Engle, Sedek, von Hecker and McIntosh2005; von Hecker et al. Reference von Hecker, Dutke and Sedek2000) we stress the specific role of limitations in mental model construction in cognitive psychopathology, especially in subclinical depression.

There are close parallels between aspects of cognitive functioning in depression and the state resulting from pre-exposure to uncontrollability. In line with the cognitive exhaustion model (Sedek & Kofta Reference Sedek and Kofta1990; von Hecker & Sedek Reference von Hecker and Sedek1999), we assume that some of the cognitive impairments observed in depression can be explained by experiences of unsolvable situations that lead to uncertainty. Such experiences may stem from past, irreversible life events, from subsequent rumination, or from counterfactual thinking. We hypothesize that uncontrollability and, in particular, ruminating thoughts about uncontrollable conditions, can lead to a depletion of those cognitive resources that support flexible, constructive thinking. Extended rumination by a victim of trauma, for example, may lead to cognitive states that impair building new cognitive models necessary for optimal functioning. Although constructive thinking may be initiated by depressive individuals, this cognitive limitation will impair the quality of new, integrative constructions or mental models related to a particular episode, a class of situations, or in more severe cases, about numerous aspects of life. Further, this may cause broader deficits given the central role of mental model construction for cognition in general (see Brewer Reference Brewer and Morris1987; Garnham Reference Garnham and Conway1997; Greeno Reference Greeno, Klahr and Kotovsky1989; Holland et al. Reference Holland, Holyoak, Nisbett and Thagard1986; Johnson-Laird Reference Johnson-Laird, Vega, Intons-Peterson, Johnson-Laird, Denis and Marschark1996).

Considering only memory processes provides an incomplete picture of cognitive targets for therapeutic change; there is compelling evidence for cognitive limitations in depression that go beyond just memory performance (Sedek et al. Reference Sedek, Brzezicka, von Hecker, Gruszka, Matthews and Szymura2010; von Hecker & Sedek Reference von Hecker and Sedek1999; von Hecker et al. Reference von Hecker, Sedek and Brzezicka2013). Indeed, depressed participants demonstrate these limitations across various paradigms tapping mental model construction: (a) mental models of interpersonal sentiment relations (social cliques models); (b) linear order reasoning (mental arrays); (c) evaluation of categorical syllogisms (mental models of logical relations); (d) situation models (inferences about the meaning of written text). Of these, we shall discuss (a) and (b) in greater detail.

Regarding (a), depressed individuals often exhibit compromised interpersonal behavior (e.g., Gotlib & Hammen Reference Gotlib and Hammen1992). Thus, we (von Hecker & Sedek Reference von Hecker and Sedek1999) studied how mental models of sentiment patterns are constructed, a crucial component of understanding one's social environment that might be affected by depression. (Participants were presented with series of pairwise sentiment relations (e.g., “Tom and Bill like each other,” “Tom and Joe dislike each other”) such that the complete set of relations formed subsets of people who like each other within cliques whilst disliking people in other cliques. Amongst all relations, a few diagnostic ones would always determine the actual number of cliques. Although depressed individuals did notice the diagnostic value of these particular relations, they were less accurate than non-depressed individuals in determining the number of cliques involved. We interpret this as a demonstration of the difficulties depressed people have with the construction of adequate social mental models (von Hecker & Sedek Reference von Hecker and Sedek1999, Experiments 2 and 3). They remembered the key elements, but they could not generate a mental model based on that information.

Regarding (b), we studied the symbolic distance effect (SDE; see Leth-Steensen & Marley Reference Leth-Steensen and Marley2000), the phenomenon that if people learn bits of information such as “Tom is older than Harry,” “Harry is older than Jack,” and “Jack is older than Bill,” they respond quicker and more accurately when later asked about the older one in pairs of persons wider apart in the ordered sequence (e.g., Tom and Bill) as compared with narrower pairs (e.g., Tom and Harry). We (Sedek & von Hecker Reference Sedek and von Hecker2004) found this effect reversed in depressed participants. Given that the SDE follows on the basis of discriminability assumptions (Holyoak & Patterson Reference Holyoak and Patterson1981) when people construct an integrated linear model of the order information (e.g., Tom>Harry>Jack>Bill), we think that depressed individuals may not readily construct such models but rather rely on the original piecemeal information when responding. Overall, mental models are a prime vehicle for individuals to determine their perspective in the world and in social contexts (Garnham Reference Garnham and Conway1997; Holland et al. Reference Holland, Holyoak, Nisbett and Thagard1986; Johnson-Laird Reference Johnson-Laird, Vega, Intons-Peterson, Johnson-Laird, Denis and Marschark1996; von Hecker et al. Reference von Hecker, Crockett, Hummert and Kemper1996) such that therapeutic intervention at this point seems essential.

Based on the above perspective and findings, we suggest that a crucial aspect of therapeutic change when dealing with depression (related to traumatic stress and other forms of emotional disturbances) may be to re-strengthen the ability to construct mental models, especially in the social domain. Concerning the therapeutic approaches to the above disturbances we also think that Lane et al.'s term “integrative memory structure” should be complemented by “construction of mental models.” Focusing on the creation of new mental models, especially for disorders such as depression, may be more consistent with the benefits seen from approaches, such as cognitive behavioral therapy, that deal with developing functional understandings and responses to current events in contrast to adjusting or understanding prior events.

Finally, we concur with Lane et al. on the importance of looking at cognitive processes as leverage points for therapeutic intervention. Cognitive processes are critical to how the internal and external world interact. We believe that as much as Lane et al. are right to stress the importance of interactions between emotion and memory content as a vantage point for therapeutic intervention, considering interactions between emotions and cognitive procedures is another useful vantage point. Moreover, our specific findings in depression underscore the importance of considering how there may be different foci for different disorders. This broader cognitive approach may have major relevance for future directions in developing therapeutic strategies.