In this commentary, we will demonstrate the usefulness of network approaches to the development of a neurobiological framework for resilience research, by relating to two points discussed in Kalisch et al.'s paper. First, we show that the paradigm shift to a transdiagnostic approach, emphasized by them, is prompted in light of recent brain network studies. Second, we propose that the network perspective may serve as a unifying framework to various resilience mechanisms, including the appraisal mechanism suggested by Kalisch et al.
Recent years have seen a surge of neuroscientific studies applying network approaches to imaging and electrophysiological data, offering a novel perspective to study brains, both in health and in psychopathology. These studies emphasize the use of graph theoretical tools to investigate brain networks. These tools characterize not only the network's architecture and possible physical dynamics but also its emergent properties, including how it copes with stressors (Sporns Reference Sporns2011).
Converging evidence indicates that healthy brains self-organize toward so-called small-world networks. A small-world architecture enables an optimal balance between local (segregation) and global (integration) structural characteristics, which is essential for global and fast information transmission (Fekete et al. Reference Fekete, Beacher, Cha, Rubin and Mujica-Parodi2014).
Relevant to the field of resilience research is the fact that the architecture of a network affects not only its function, but also its robustness to perturbations (Kaiser et al. Reference Kaiser, Martin, Andras and Young2007). With regard to the brain, network analysis of both structural and functional data suggests that brains are highly robust systems (Kaiser et al. Reference Kaiser, Martin, Andras and Young2007). When abnormalities in network metrics occur, however, leading to deviation from healthy architecture, general network failure can result (Sporns Reference Sporns2011). Such network failure can affect network robustness as well as efficient information transfer; it also can cause deficits in the access, engagement, and disengagement of large-scale networks supporting cognition and behavior (Dosenbach et al. Reference Dosenbach, Fair, Cohen, Schlaggar and Petersen2008). This understanding has led us recently to suggest that network architecture is directly associated with mental resilience (Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013; Levit-Binnun & Golland Reference Levit-Binnun and Golland2011).
With the understanding how deviations from healthy architecture can have profound consequences on brain function and robustness, we now can proceed to explain why the network perspective prompts a shift to a transdiagnostic approach. According to the network perspective, brain diseases are cases of deviation from healthy architecture (Menon Reference Menon2011). Indeed, deviations from healthy architecture have been found across a wide range of psychopathologies, including developmental (e.g., attention deficit hyperactivity disorder and autism), psychiatric (e.g., schizophrenia, major depression, posttraumatic stress disorder, and obsessive-compulsive disorder), and neurological disorders (e.g., Alzheimer's and other dementias) (for a review, see Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013).
Even more interesting, these deviations from healthy network architecture are evident already in individuals who are at risk for psychopathology but who do not present clinical signs. Hence, neonates at genetic risk for schizophrenia, individuals presenting autistic traits or carrying an autism risk gene, individuals with familial risk for ADHD, children at risk for anxiety and depression, and individuals with mild cognitive impairments all present networks that deviate from healthy architecture (for a review, see Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013).
That deviations from healthy network architecture are found across different psychopathologies and in at-risk populations supports the need for the transdiagnostic approach in resilience research that Kalisch et al. describe. Moreover, this suggests that aberrations in network metrics may be indicators of global vulnerability-conducive traits (Fekete et al. Reference Fekete, Beacher, Cha, Rubin and Mujica-Parodi2014; Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013). Indeed, we recently showed that psychopathological networks (schizophrenia) that deviate from healthy structure respond abnormally to a controlled perturbation induced by transcranial magnetic stimulation pulses (Arzouan et al. Reference Arzouan, Moses, Peled and Levit-Binnun2014).
The network perspective also supports the attempt made by Kalisch et al. to develop a unifying framework for the study of general resilience mechanisms. Indeed, various internal and external resilience factors (e.g., cognitive abilities, personality traits, (epi)genetics, age, sex, spirituality, social support, and socioeconomic status – all factors mentioned in the target article) can be framed within the network perspective. For example, individual differences in cognitive ability have been linked with variations in network metrics (van den Heuvel et al. Reference van den Heuvel, Stam, Kahn and Hulshoff Pol2009). Personality traits, such as low effortful control, have been associated with compromised small-world connectivity (Fekete et al. Reference Fekete, Beacher, Cha, Rubin and Mujica-Parodi2014). Genetics, age, and gender are all factors that have been found to influence network architecture (Gong et al. Reference Gong, Rosa-Neto, Carbonell, Chen, He and Evans2009; Stam & van Straaten Reference Stam and van Straaten2012). Other resilience-related factors (e.g., social support, socioeconomic status, and spiritual activities) have not been directly studied within a network framework. As most of them have been related to distributed changes in the brain (Hölzel et al. Reference Hölzel, Carmody, Vangel, Congleton, Yerramsetti, Gard and Lazar2011; Kanai et al. Reference Kanai, Bahrami, Roylance and Rees2012; McEwen & Gianaros Reference McEwen and Gianaros2010), however, one may hypothesize that these changes also would manifest at the level of network metrics and global architecture.
Such framing of resilience-related factors within the network perspective also can be extended to the specific appraisal mechanism suggested by Kalisch et al. We suggest that a healthy and optimal network architecture is a necessary neurobiological condition for intact appraisal mechanisms. Abnormalities in network metrics (whether inherited or developed) lead both to the disruption of dynamic balance and to deficient information integration and transfer. This can lead to irregularities in low-level functions such as sensory, motor, and regulatory processes (Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013; Levit-Binnun & Golland Reference Levit-Binnun and Golland2011), which can, in turn, lead to abnormalities in the way one evaluates and reacts to stressors. Moreover, as the authors state, appraisal is not a single process but probably multiple, dynamic, and interactive operations. Abnormalities in network architecture may cause deficits in the access, engagement, and disengagement of these complex operations (Dosenbach et al. Reference Dosenbach, Fair, Cohen, Schlaggar and Petersen2008) leading to deficient appraisal mechanisms.
In sum, we suggest that a network perspective supports a transdiagnostic approach to resilience and can contribute to the development of a unifying framework for studying global resilience mechanisms. Notably, a direct link between network architecture and mental resilience remains to be demonstrated. The accumulating evidence nonetheless suggests that network approaches are highly relevant to the search of a neurobiological framework for the study of resilience.
In this commentary, we will demonstrate the usefulness of network approaches to the development of a neurobiological framework for resilience research, by relating to two points discussed in Kalisch et al.'s paper. First, we show that the paradigm shift to a transdiagnostic approach, emphasized by them, is prompted in light of recent brain network studies. Second, we propose that the network perspective may serve as a unifying framework to various resilience mechanisms, including the appraisal mechanism suggested by Kalisch et al.
Recent years have seen a surge of neuroscientific studies applying network approaches to imaging and electrophysiological data, offering a novel perspective to study brains, both in health and in psychopathology. These studies emphasize the use of graph theoretical tools to investigate brain networks. These tools characterize not only the network's architecture and possible physical dynamics but also its emergent properties, including how it copes with stressors (Sporns Reference Sporns2011).
Converging evidence indicates that healthy brains self-organize toward so-called small-world networks. A small-world architecture enables an optimal balance between local (segregation) and global (integration) structural characteristics, which is essential for global and fast information transmission (Fekete et al. Reference Fekete, Beacher, Cha, Rubin and Mujica-Parodi2014).
Relevant to the field of resilience research is the fact that the architecture of a network affects not only its function, but also its robustness to perturbations (Kaiser et al. Reference Kaiser, Martin, Andras and Young2007). With regard to the brain, network analysis of both structural and functional data suggests that brains are highly robust systems (Kaiser et al. Reference Kaiser, Martin, Andras and Young2007). When abnormalities in network metrics occur, however, leading to deviation from healthy architecture, general network failure can result (Sporns Reference Sporns2011). Such network failure can affect network robustness as well as efficient information transfer; it also can cause deficits in the access, engagement, and disengagement of large-scale networks supporting cognition and behavior (Dosenbach et al. Reference Dosenbach, Fair, Cohen, Schlaggar and Petersen2008). This understanding has led us recently to suggest that network architecture is directly associated with mental resilience (Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013; Levit-Binnun & Golland Reference Levit-Binnun and Golland2011).
With the understanding how deviations from healthy architecture can have profound consequences on brain function and robustness, we now can proceed to explain why the network perspective prompts a shift to a transdiagnostic approach. According to the network perspective, brain diseases are cases of deviation from healthy architecture (Menon Reference Menon2011). Indeed, deviations from healthy architecture have been found across a wide range of psychopathologies, including developmental (e.g., attention deficit hyperactivity disorder and autism), psychiatric (e.g., schizophrenia, major depression, posttraumatic stress disorder, and obsessive-compulsive disorder), and neurological disorders (e.g., Alzheimer's and other dementias) (for a review, see Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013).
Even more interesting, these deviations from healthy network architecture are evident already in individuals who are at risk for psychopathology but who do not present clinical signs. Hence, neonates at genetic risk for schizophrenia, individuals presenting autistic traits or carrying an autism risk gene, individuals with familial risk for ADHD, children at risk for anxiety and depression, and individuals with mild cognitive impairments all present networks that deviate from healthy architecture (for a review, see Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013).
That deviations from healthy network architecture are found across different psychopathologies and in at-risk populations supports the need for the transdiagnostic approach in resilience research that Kalisch et al. describe. Moreover, this suggests that aberrations in network metrics may be indicators of global vulnerability-conducive traits (Fekete et al. Reference Fekete, Beacher, Cha, Rubin and Mujica-Parodi2014; Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013). Indeed, we recently showed that psychopathological networks (schizophrenia) that deviate from healthy structure respond abnormally to a controlled perturbation induced by transcranial magnetic stimulation pulses (Arzouan et al. Reference Arzouan, Moses, Peled and Levit-Binnun2014).
The network perspective also supports the attempt made by Kalisch et al. to develop a unifying framework for the study of general resilience mechanisms. Indeed, various internal and external resilience factors (e.g., cognitive abilities, personality traits, (epi)genetics, age, sex, spirituality, social support, and socioeconomic status – all factors mentioned in the target article) can be framed within the network perspective. For example, individual differences in cognitive ability have been linked with variations in network metrics (van den Heuvel et al. Reference van den Heuvel, Stam, Kahn and Hulshoff Pol2009). Personality traits, such as low effortful control, have been associated with compromised small-world connectivity (Fekete et al. Reference Fekete, Beacher, Cha, Rubin and Mujica-Parodi2014). Genetics, age, and gender are all factors that have been found to influence network architecture (Gong et al. Reference Gong, Rosa-Neto, Carbonell, Chen, He and Evans2009; Stam & van Straaten Reference Stam and van Straaten2012). Other resilience-related factors (e.g., social support, socioeconomic status, and spiritual activities) have not been directly studied within a network framework. As most of them have been related to distributed changes in the brain (Hölzel et al. Reference Hölzel, Carmody, Vangel, Congleton, Yerramsetti, Gard and Lazar2011; Kanai et al. Reference Kanai, Bahrami, Roylance and Rees2012; McEwen & Gianaros Reference McEwen and Gianaros2010), however, one may hypothesize that these changes also would manifest at the level of network metrics and global architecture.
Such framing of resilience-related factors within the network perspective also can be extended to the specific appraisal mechanism suggested by Kalisch et al. We suggest that a healthy and optimal network architecture is a necessary neurobiological condition for intact appraisal mechanisms. Abnormalities in network metrics (whether inherited or developed) lead both to the disruption of dynamic balance and to deficient information integration and transfer. This can lead to irregularities in low-level functions such as sensory, motor, and regulatory processes (Levit-Binnun et al. Reference Levit-Binnun, Davidovitch and Golland2013; Levit-Binnun & Golland Reference Levit-Binnun and Golland2011), which can, in turn, lead to abnormalities in the way one evaluates and reacts to stressors. Moreover, as the authors state, appraisal is not a single process but probably multiple, dynamic, and interactive operations. Abnormalities in network architecture may cause deficits in the access, engagement, and disengagement of these complex operations (Dosenbach et al. Reference Dosenbach, Fair, Cohen, Schlaggar and Petersen2008) leading to deficient appraisal mechanisms.
In sum, we suggest that a network perspective supports a transdiagnostic approach to resilience and can contribute to the development of a unifying framework for studying global resilience mechanisms. Notably, a direct link between network architecture and mental resilience remains to be demonstrated. The accumulating evidence nonetheless suggests that network approaches are highly relevant to the search of a neurobiological framework for the study of resilience.