Limits...
A Complex Systems Approach to Causal Discovery in Psychiatry.

Saxe GN, Statnikov A, Fenyo D, Ren J, Li Z, Prasad M, Wall D, Bergman N, Briggs EC, Aliferis C - PLoS ONE (2016)

Bottom Line: This causal network had well-defined adaptive properties and a set of variables was found that disproportionally contributed to these properties.Modeling the removal of these variables resulted in significant loss of adaptive properties.The novel approach yields both fine-grain (micro) and high-level (macro) insights and thus represents a promising approach for complex systems-oriented research in psychiatry.

View Article: PubMed Central - PubMed

Affiliation: Department of Child and Adolescent Psychiatry, New York University School of Medicine, New York, New York, United States of America.

ABSTRACT
Conventional research methodologies and data analytic approaches in psychiatric research are unable to reliably infer causal relations without experimental designs, or to make inferences about the functional properties of the complex systems in which psychiatric disorders are embedded. This article describes a series of studies to validate a novel hybrid computational approach--the Complex Systems-Causal Network (CS-CN) method-designed to integrate causal discovery within a complex systems framework for psychiatric research. The CS-CN method was first applied to an existing dataset on psychopathology in 163 children hospitalized with injuries (validation study). Next, it was applied to a much larger dataset of traumatized children (replication study). Finally, the CS-CN method was applied in a controlled experiment using a 'gold standard' dataset for causal discovery and compared with other methods for accurately detecting causal variables (resimulation controlled experiment). The CS-CN method successfully detected a causal network of 111 variables and 167 bivariate relations in the initial validation study. This causal network had well-defined adaptive properties and a set of variables was found that disproportionally contributed to these properties. Modeling the removal of these variables resulted in significant loss of adaptive properties. The CS-CN method was successfully applied in the replication study and performed better than traditional statistical methods, and similarly to state-of-the-art causal discovery algorithms in the causal detection experiment. The CS-CN method was validated, replicated, and yielded both novel and previously validated findings related to risk factors and potential treatments of psychiatric disorders. The novel approach yields both fine-grain (micro) and high-level (macro) insights and thus represents a promising approach for complex systems-oriented research in psychiatry.

Show MeSH

Related in: MedlinePlus

Integrity of CHIDS Causal Network Following Challenge.The proportion of nodes in the largest network component by sequential removal of 15 nodes at random vs. by BC rank.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4814084&req=5

pone.0151174.g006: Integrity of CHIDS Causal Network Following Challenge.The proportion of nodes in the largest network component by sequential removal of 15 nodes at random vs. by BC rank.

Mentions: The final step in our approach is to test the adaptive nature of the network by modeling how it behaves when stressed by the sequential removal of 15 nodes at random vs. by BC rank order. In Fig 6, this differential response to challenge is shown dynamically based on the network’s capacity to maintain its largest component, a strong indicator of robustness. We examine the proportion of nodes that remain in the largest component of the network following sequential removal by rank order based on BC rank or by random number generator (with recalculation of rank at each step) [12]. Each step of the random node removal calculates the mean of the remaining proportion of nodes in the largest component from 1000 trials. As can be seen, random sequential removal of 15 nodes has negligible impact on the network. The CHIDS network maintains close to 100% of its integrity throughout this sequential challenge. On the other hand, the sequential removal of nodes by order of BC rank leads to a considerable loss of network integrity. By the end of this 15 node sequential ‘attack’ the CHIDS network has lost almost 80% of its nodes. As can be seen, its integrity is maintained through the sequential removal of its four highest BC ranked nodes, and then the network quickly loses its integrity with more sequential ‘attacks.’


A Complex Systems Approach to Causal Discovery in Psychiatry.

Saxe GN, Statnikov A, Fenyo D, Ren J, Li Z, Prasad M, Wall D, Bergman N, Briggs EC, Aliferis C - PLoS ONE (2016)

Integrity of CHIDS Causal Network Following Challenge.The proportion of nodes in the largest network component by sequential removal of 15 nodes at random vs. by BC rank.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4814084&req=5

pone.0151174.g006: Integrity of CHIDS Causal Network Following Challenge.The proportion of nodes in the largest network component by sequential removal of 15 nodes at random vs. by BC rank.
Mentions: The final step in our approach is to test the adaptive nature of the network by modeling how it behaves when stressed by the sequential removal of 15 nodes at random vs. by BC rank order. In Fig 6, this differential response to challenge is shown dynamically based on the network’s capacity to maintain its largest component, a strong indicator of robustness. We examine the proportion of nodes that remain in the largest component of the network following sequential removal by rank order based on BC rank or by random number generator (with recalculation of rank at each step) [12]. Each step of the random node removal calculates the mean of the remaining proportion of nodes in the largest component from 1000 trials. As can be seen, random sequential removal of 15 nodes has negligible impact on the network. The CHIDS network maintains close to 100% of its integrity throughout this sequential challenge. On the other hand, the sequential removal of nodes by order of BC rank leads to a considerable loss of network integrity. By the end of this 15 node sequential ‘attack’ the CHIDS network has lost almost 80% of its nodes. As can be seen, its integrity is maintained through the sequential removal of its four highest BC ranked nodes, and then the network quickly loses its integrity with more sequential ‘attacks.’

Bottom Line: This causal network had well-defined adaptive properties and a set of variables was found that disproportionally contributed to these properties.Modeling the removal of these variables resulted in significant loss of adaptive properties.The novel approach yields both fine-grain (micro) and high-level (macro) insights and thus represents a promising approach for complex systems-oriented research in psychiatry.

View Article: PubMed Central - PubMed

Affiliation: Department of Child and Adolescent Psychiatry, New York University School of Medicine, New York, New York, United States of America.

ABSTRACT
Conventional research methodologies and data analytic approaches in psychiatric research are unable to reliably infer causal relations without experimental designs, or to make inferences about the functional properties of the complex systems in which psychiatric disorders are embedded. This article describes a series of studies to validate a novel hybrid computational approach--the Complex Systems-Causal Network (CS-CN) method-designed to integrate causal discovery within a complex systems framework for psychiatric research. The CS-CN method was first applied to an existing dataset on psychopathology in 163 children hospitalized with injuries (validation study). Next, it was applied to a much larger dataset of traumatized children (replication study). Finally, the CS-CN method was applied in a controlled experiment using a 'gold standard' dataset for causal discovery and compared with other methods for accurately detecting causal variables (resimulation controlled experiment). The CS-CN method successfully detected a causal network of 111 variables and 167 bivariate relations in the initial validation study. This causal network had well-defined adaptive properties and a set of variables was found that disproportionally contributed to these properties. Modeling the removal of these variables resulted in significant loss of adaptive properties. The CS-CN method was successfully applied in the replication study and performed better than traditional statistical methods, and similarly to state-of-the-art causal discovery algorithms in the causal detection experiment. The CS-CN method was validated, replicated, and yielded both novel and previously validated findings related to risk factors and potential treatments of psychiatric disorders. The novel approach yields both fine-grain (micro) and high-level (macro) insights and thus represents a promising approach for complex systems-oriented research in psychiatry.

Show MeSH
Related in: MedlinePlus