The problems identified by Borsboom et al. with regard to reductionism in neuropsychiatry are evident in the widespread use of animal models of mental disorders in modern neuroscience. Neuroscience research focused on psychiatric disorders is dominated by studies using a small number of species – predominantly, artificially housed inbred strains of laboratory mice. This work uses selective breeding, genetic engineering (to produce transgenic lines or mutant knockouts), targeted lesioning of the brain, or manipulations of the environment to recapitulate the plausible causative factor(s) thought to underlie a given diagnosis, or at least the neural or behavioral pathologies which characterize the human disorder.
Unfortunately, these animal models often are found to have weak correspondence to the phenomenology of the neuropsychiatric disorder in question (weak validation), and drugs developed using these models often have limited efficacy (poor predictive validity) (Markou et al. Reference Markou, Chiamulera, Geyer, Tricklebank and Steckler2009; Nestler & Hyman Reference Nestler and Hyman2010). For example, despite hundreds of supposed mouse models of autism spectrum disorder (ASD), no pharmacological interventions have yet been found that markedly improve ASD's characteristic deficits in either social interaction or repetitive behavior (Kazdoba et al. Reference Kazdoba, Leach, Yang, Silverman, Solomon, Crawley, Robbins and Sahakian2016; Varghese et al. Reference Varghese, Keshav, Jacot-Descombes, Warda, Wicinski, Dickstein, Harony-Nicolas, de Rubeis, Drapeau, Buxbaum and Hof2017). Furthermore, several widely used animal models of depression (e.g., tests of behavioral despair or learned helplessness, such as forced-swim or tail-suspension tests) map poorly to the pathologies of chronic depression, which also include symptoms such as anhedonia, disruptions of sleep, and changes to psychomotor behavior. And, of course, animal models will never be able to recapitulate symptoms that are central to many neuropsychiatric disorders: the content of mental states. It is hard to imagine a mouse that experiences rumination, guilt, shame, or existential ennui.
Unfortunately, the reductive emphasis on rodent models of human mental disorders in neuroscience may be hindering the development of more safe and effective interventions for psychiatric patients. While much can be learned from neuroscience research which attempts to understand the genetic, anatomical, and molecular mechanisms underlying mental disorders, more phenomenologically and taxonomically broad efforts targeted at understanding what Borsboom and colleagues call “network structures” will provide key additional insights. Comparative research that focuses on diversity and variation between species, instead of merely attempting to phenocopy a human disorder in a single – albeit convenient – species, will be necessary to build a comprehensive understanding of the general principles of brain organization and development (Striedter et al. Reference Striedter, Belgard, Chen, Davis, Finlay, Güntürkün, Hale, Harris, Hecht, Hof, Hofmann, Holland, Iwaniuk, Jarvis, Karten, Katz, Kristan, Macagno, Mitra, Moroz, Preuss, Ragsdale, Sherwood, Stevens, Stüttgen, Tsumoto and Wilczynski2014).
The problem is that reductive research efforts are prioritized over more broad research efforts by funding agencies, journals editors, hiring committees, and the popular press. Incentives in neuroscience often reward technologically complex experiments focused on dissecting neural circuits, but not careful behavioral observation, long-term studies that investigate the developmental trajectories of behavior, or work that seeks to understand the environmental and evolutionary context in which any given behavior functions (see Krakauer et al. [Reference Krakauer, Ghazanfar, Gomez-Marin, MacIver and Poeppel2017] for a review).
A broader, holistic, more evolutionarily grounded approach may ultimately provide critical insights into the brain circuits, developmental processes, genetic mechanisms, and environmental factors which contribute to the “massively multifactorial system networks” that go awry in mental disorders. Indeed, some of the most exciting discoveries about the functioning of the brain have come from long-term neuroethological studies of “non-model” organisms, such as pair bonding in prairie voles, vocal learning in songbirds, spatial attention in owls, and social plasticity in cichlid fish (Knudsen Reference Knudsen2011; Maruska & Fernald Reference Maruska and Fernald2018; McGraw & Young Reference McGraw and Young2010; Pfenning et al. Reference Pfenning, Hara, Whitney, Rivas, Wang, Roulhac, Howard, Wirthlin, Lovell, Ganapathy, Mountcastle, Moseley, Thompson, Soderblom, Iriki, Kato, Gilbert, Zhang, Bakken, Bongaarts, Bernard, Lein, Mello, Hartemink and Jarvis2014). This work is especially important because many key behaviors at the core of some neuropsychiatric disorders have no parallel in the behavioral repertoire of a rodent (e.g., language-learning deficits in ASD).
Additionally, greater emphasis should be placed on understanding the specific mechanisms by which the environment and experience (stress, trauma, interactions with caregivers, etc.) are translated into changes in the brain. Learning and responding to the environment is what brains do, and brains are profoundly shaped by experience at every stage of development. In addition to genetic correlates, the majority of neuropsychiatric disorders have profoundly important, yet woefully understudied, environmental etiologies. As such, researchers should not shy away from performing long-term developmental experiments to understand these mechanisms. Fortunately, there is a renewed interest in the neurobiological mechanisms of plasticity, including the genes, molecules, and epigenetic influences which regulate the sensitivity of organisms to environmental conditions (Baran Reference Baran2017; Caspi et al. Reference Caspi, Hariri, Holmes, Uher and Moffitt2010; Meaney Reference Meaney, Tolan and Leventhal2017). However, much of this work is still in its infancy.
Reductive approaches in neuroscience research have led to an extreme focus on trying to find biologically based interventions (i.e., drug development), despite the fact that we already know that behavioral and environmental interventions are critical components in the effective treatment of many human mental disorders. Taking a network structure approach suggests that we should both include a greater diversity of organisms and behaviors in neuroscience research, as well as study complex interactions between multiple factors at multiple levels of analysis. Borsboom et al. have identified a conceptual flaw at the heart of neuropsychiatric research; preclinical neuroscience researchers would do well to heed the warning.
The problems identified by Borsboom et al. with regard to reductionism in neuropsychiatry are evident in the widespread use of animal models of mental disorders in modern neuroscience. Neuroscience research focused on psychiatric disorders is dominated by studies using a small number of species – predominantly, artificially housed inbred strains of laboratory mice. This work uses selective breeding, genetic engineering (to produce transgenic lines or mutant knockouts), targeted lesioning of the brain, or manipulations of the environment to recapitulate the plausible causative factor(s) thought to underlie a given diagnosis, or at least the neural or behavioral pathologies which characterize the human disorder.
Unfortunately, these animal models often are found to have weak correspondence to the phenomenology of the neuropsychiatric disorder in question (weak validation), and drugs developed using these models often have limited efficacy (poor predictive validity) (Markou et al. Reference Markou, Chiamulera, Geyer, Tricklebank and Steckler2009; Nestler & Hyman Reference Nestler and Hyman2010). For example, despite hundreds of supposed mouse models of autism spectrum disorder (ASD), no pharmacological interventions have yet been found that markedly improve ASD's characteristic deficits in either social interaction or repetitive behavior (Kazdoba et al. Reference Kazdoba, Leach, Yang, Silverman, Solomon, Crawley, Robbins and Sahakian2016; Varghese et al. Reference Varghese, Keshav, Jacot-Descombes, Warda, Wicinski, Dickstein, Harony-Nicolas, de Rubeis, Drapeau, Buxbaum and Hof2017). Furthermore, several widely used animal models of depression (e.g., tests of behavioral despair or learned helplessness, such as forced-swim or tail-suspension tests) map poorly to the pathologies of chronic depression, which also include symptoms such as anhedonia, disruptions of sleep, and changes to psychomotor behavior. And, of course, animal models will never be able to recapitulate symptoms that are central to many neuropsychiatric disorders: the content of mental states. It is hard to imagine a mouse that experiences rumination, guilt, shame, or existential ennui.
Unfortunately, the reductive emphasis on rodent models of human mental disorders in neuroscience may be hindering the development of more safe and effective interventions for psychiatric patients. While much can be learned from neuroscience research which attempts to understand the genetic, anatomical, and molecular mechanisms underlying mental disorders, more phenomenologically and taxonomically broad efforts targeted at understanding what Borsboom and colleagues call “network structures” will provide key additional insights. Comparative research that focuses on diversity and variation between species, instead of merely attempting to phenocopy a human disorder in a single – albeit convenient – species, will be necessary to build a comprehensive understanding of the general principles of brain organization and development (Striedter et al. Reference Striedter, Belgard, Chen, Davis, Finlay, Güntürkün, Hale, Harris, Hecht, Hof, Hofmann, Holland, Iwaniuk, Jarvis, Karten, Katz, Kristan, Macagno, Mitra, Moroz, Preuss, Ragsdale, Sherwood, Stevens, Stüttgen, Tsumoto and Wilczynski2014).
The problem is that reductive research efforts are prioritized over more broad research efforts by funding agencies, journals editors, hiring committees, and the popular press. Incentives in neuroscience often reward technologically complex experiments focused on dissecting neural circuits, but not careful behavioral observation, long-term studies that investigate the developmental trajectories of behavior, or work that seeks to understand the environmental and evolutionary context in which any given behavior functions (see Krakauer et al. [Reference Krakauer, Ghazanfar, Gomez-Marin, MacIver and Poeppel2017] for a review).
A broader, holistic, more evolutionarily grounded approach may ultimately provide critical insights into the brain circuits, developmental processes, genetic mechanisms, and environmental factors which contribute to the “massively multifactorial system networks” that go awry in mental disorders. Indeed, some of the most exciting discoveries about the functioning of the brain have come from long-term neuroethological studies of “non-model” organisms, such as pair bonding in prairie voles, vocal learning in songbirds, spatial attention in owls, and social plasticity in cichlid fish (Knudsen Reference Knudsen2011; Maruska & Fernald Reference Maruska and Fernald2018; McGraw & Young Reference McGraw and Young2010; Pfenning et al. Reference Pfenning, Hara, Whitney, Rivas, Wang, Roulhac, Howard, Wirthlin, Lovell, Ganapathy, Mountcastle, Moseley, Thompson, Soderblom, Iriki, Kato, Gilbert, Zhang, Bakken, Bongaarts, Bernard, Lein, Mello, Hartemink and Jarvis2014). This work is especially important because many key behaviors at the core of some neuropsychiatric disorders have no parallel in the behavioral repertoire of a rodent (e.g., language-learning deficits in ASD).
Additionally, greater emphasis should be placed on understanding the specific mechanisms by which the environment and experience (stress, trauma, interactions with caregivers, etc.) are translated into changes in the brain. Learning and responding to the environment is what brains do, and brains are profoundly shaped by experience at every stage of development. In addition to genetic correlates, the majority of neuropsychiatric disorders have profoundly important, yet woefully understudied, environmental etiologies. As such, researchers should not shy away from performing long-term developmental experiments to understand these mechanisms. Fortunately, there is a renewed interest in the neurobiological mechanisms of plasticity, including the genes, molecules, and epigenetic influences which regulate the sensitivity of organisms to environmental conditions (Baran Reference Baran2017; Caspi et al. Reference Caspi, Hariri, Holmes, Uher and Moffitt2010; Meaney Reference Meaney, Tolan and Leventhal2017). However, much of this work is still in its infancy.
Reductive approaches in neuroscience research have led to an extreme focus on trying to find biologically based interventions (i.e., drug development), despite the fact that we already know that behavioral and environmental interventions are critical components in the effective treatment of many human mental disorders. Taking a network structure approach suggests that we should both include a greater diversity of organisms and behaviors in neuroscience research, as well as study complex interactions between multiple factors at multiple levels of analysis. Borsboom et al. have identified a conceptual flaw at the heart of neuropsychiatric research; preclinical neuroscience researchers would do well to heed the warning.