What is this article about?

There are a lot of therapies with the goal to treat mental disorders. However, there is not much known about how these treatments work, so there is not much known about the processes. This is bad, because understanding which factors facilitate and inhibit therapeutic change can guide researchers in improving treatments, relapse, and recurrence. The complex system approach is used to study change across different physical and natural systems, ranging from cells and neurons to political and economic systems. There have been attempts to use the complex systems science in psychology and psychiatry, but this is a slow process. One reason for this is the randomized controlled trial (RCT) design. RCT is important for evaluating treatment efficacy, but these treatments are often focused at one component of functioning, such as cognitions, emotions, behavior, or physiology, instead of at the multi-component pattern, as in the complex systems approach. Another barrier to the uptake of the complex systems approach is that different subdisciplines in psychology and psychotherapy use different jargon and concepts. This makes it difficult to detect common themes and principles. In this article, the authors advance the complex systems approach by presenting some basic principles in a way that is true to complexity science, and which is accessible to researchers and clinicians. The goal is to provide an integrative framework which helps to translate concepts into a common language and to provide a structure for conceptualizing and studying different treatments and clinical problems.

What are the general principles in complex adaptive systems?

Pattern formation and attractors

A dynamic system is defined as a set of interconnected elements (connected to each other) which evolve over time and self-organize into higher-order functional units, called attractor states. Self-organization is defined as the process by which lower-order processes interact and higher-order patterns emerge and influence lower-order processes in a top-down way. These attractor states predict behavior. When the states are changing, the attractor states ‘attract’ the behavior back to this. Attractor states that are well-established have strong interconnected elements, with reinforcing and inhibiting feedback loops. These feedback loops can decrease or increase the probability of activation over time and contexts. These attractor states do not change easily: there need to be significant disturbance before the patterns are broken. However, less developed attractors are more easily changed.

System change: tipping points and nonlinear transitions

Complex systems can adapt to defend against challenges. There can be deterministic (purposeful, causal) and stochastic (naturally occurring random events, fluctuations, noise) forces which can affect the complex systems. The chance of transitioning from one attractor to another depends on the strength of that attractor, the type of perturbation, the parameters that control system organization, and the strength of alternate attractors. Change can be gradual, but when the parameters reach a ‘tipping point’, the dominant state can shift suddenly. This type of change is often abrupt, with periods of turbulence, in which attractors destabilize and create the potential for phase or order transitions. During these transitions, systems can reorganize into new patterns of functioning, and this may lead one’s health to change from healthy to diseased. It can be important to know when these transitions will occur. For example, early warning signs of symptom exacerbation or transition to disease can inform medici about treatment decisions and can even be a matter of life and death. Therefore, scientists have identified early warning signs that occur often. Two of these early warning signs are critical fluctuations, and critical slowing. Transitions can involve movement from healthy to maladaptive states, but also movement in an opposite direction. For example, critical slowing precedes recovery with interventions.

Patterns are weak, until they are strengthened and stabilized through repeated activation across contexts and by feedback loops. There can also be a period of vacillation or ‘flickering’ between attractors. When a new attractor strengthens, it can compete with a pre-existing attractor to prevent a return to that attractor. For example, when someone is trying to create a healthy habit, their previous (unhealthy) habits are strong. One needs to repeatedly engage in this new healthy habit, until this habit is sufficiently strong, and consolidated and maintained in memory. Processes in complex systems can also operate on different timescales: some move slowly, and others move quickly.

Application to psychotherapy: network destabilization and transition model

The authors developed ‘the network destabilization and transition model’ (NDT) as a framework that includes concepts and principles from complexity sciences (dynamical systems theory, synergetics, and network science) and uses this for psychotherapy research. This model was initially meant to be used in the treatment of depression, but it can be used for psychotherapy in general. The goal of the authors is to stimulate new research and to provide a framework for understanding and organizing findings from different research.

The goal of psychotherapy is to promote new learning, so to move a person from entrenched patterns of psychopathology to more flexible and functional patterns of functioning. Some researchers describe psychopathology as an attractor state with interacting elements between cognitions, emotions, behavior, and physiology. The goal of therapy is to change these patterns and processes. Tschacher and Haken also suggest to consider contextual factors, the therapeutic relationship, and environmental and random factor (stochastic), which can all influence the change process.

There are different routes in which therapeutic change can happen. It can be that more than two patterns are relevant to psychopathology. And, it is not clear whether pathological and healthy states are different networks, or rather parts of a single, large network.

Change in psychotherapy can thus happen in different forms. First, there can be minor adjustments made to maladaptive patterns. For example, harm reduction strategies reduce some negative consequences of pathological patterns. However, they do not lead to complete abolishment of attractor states. For example, think of providing clean needles to drug addicts. Other therapies are distress tolerance, mindfulness, and positive emotion activation approaches. These can change the threshold of activation and automaticity of both pathological and more functional patterns. For example, behavioral, interpersonal, cognitive reappraisal, emotion regulation, or parenting skills can all be used to: reduce feedback loops that block new information and interfere with new learning, deactivate or unhook from the pathological patterns of the attractor, and/or compensate for, or override, new patterns. Also, these strategies can be used to decrease exposure to stochastic factors, or to reduce the influence of these factors. Thus, all of these strategies work within the pathological attractor, but they do not change the attractor directly!

Another type of change refers to ‘switching’ from a pathological to a healthy attractor. To achieve this, there should be an alternative available. For example, the therapist can provide a supportive environment and therapeutic alliance, so to increase the patient’s readiness, resources, and skills to develop healthy behaviors. For example, Beck’s recovery-oriented cognitive therapy for schizophrenia. In this intervention, patients learn to switch from a ‘patient mode’ to an ‘adaptive mode’. Thus, they need to switch from a disorder-focused mode to a mode in which they create positive beliefs, aspirations, strengths, and values. Then, this adaptive mode is constantly activated and exercised to increase its accessibility and strength. Also, the positive emotion activation approaches can help to build more healthy attractors. Further, another type of change involves destabilizing the pathological attractor and developing a new, healthier attractor. This can be achieved by exposure therapy, insight-oriented therapy, and emotion-focused and cognitive restructuring techniques.

What are some important considerations?

The complex systems approach as described is important for psychotherapy research. It emphasizes the need for longitudinal data, the study of discontinuous and nonlinear change, and a focus on patterns of functioning rather than single components.

Data collection considerations

When conducting research with time-series, it is important to select the appropriate time interval, namely the time interval (sampling rate) that is most sensitive for detecting changes in the variables of interest. For example, some variables change slowly, while others change more quickly.

Breadth and duration of assessment

When researchers use microanalytic assessments, such as assessments on a timescale of minutes, the sampling rate is high, but the disadvantage is that the number of variables and the duration of assessments is low. Ideally, researchers should measure pathological patterns and symptoms over the course of therapy, between sessions, and after therapy to capture therapeutic change. Researchers could also gather passive data (activity level, exercise, sleep, social media usage).

Level of analysis

Complex systems is an individual-level approach, and psychotherapy is in contrast often a nomothetic level of group averages. However, it is important to use individual-level data, because findings from one level might not directly generalize to another level. This individual-level data allows studying the dynamics of a given person., which can help in treatments. However, it is important for science to detect patterns and principles that generalize across all people. These problems can be bridged, by identifying common indices of early warning signals across systems and sciences. Ellison and colleagues used ecological momentary assessments methods and Group Iterative Multiple Model Estimation (GIMME) and show how these levels can be combined.

What are further important considerations?

Different trajectories of change

When using the complex systems approach, trajectories of symptom change are important. In psychotherapy, a general assumption is that change is gradual and linear. However, time course data has shown that this is not always the case, and that the process can follow a nonlinear course. Changes in therapy can follow a quadratic pattern (U-shape), a cubic pattern, a saw-toothed pattern, and other nonlinear patterns. Think of relapse in addiction treatment.

Early warning signals

Increased variability and turbulence in psychotherapy is seen as potential for change according to the complex systems approach. There have not been many studies that have looked at early warning signals in psychotherapy. There are different ways to measure early warning signals. For example, using Grid-Ware, recurrence quantification analysis, or dynamic complexity. Calculating dynamic complexity can be done through R.  An important avenue for future research is to examine which early warning signals predict transition across clinical problems and treatments, and which represent therapeutic change.

Patterns and feedback loops

Thus, as noted, attractors exist of interconnected elements. Thus, patterns are the focus of study, and not a single component. Again, recurrence quantification analysis and related tools (Grid-Ware), the Synergetic Navigation System can be used to study multi-component patterns for individuals. One can also use network analysis tools, which quantify the structure, density, connectivity, and threshold of activation of patterns, and how they change over time. This network analysis depicts and measures patterns of psychopathology for a certain sample, and also personalized for a specific individual. This may guide treatment decisions and selection. However, a limitation of network analysis is that it assumes ‘stationarity’, which means that each variable, over time, demonstrates similar means, variances, and relationships with other variables and with itself.

Interplay of pathological and new patterns of learning

The complex systems approach thus suggests that new attractors can develop, and can build in strength so that they can compete with or inhibit old attractors. Modern cognitive development and learning theories also suggest that psychotherapy promotes new learning, through establishment of new patterns. However, there has not been much research conducted into how this old-new attractor competition takes place.

What can be concluded?

Thus, the NDT model for therapeutic change is a conceptual framework. The goal is to identify and translate concepts from subdisciplines of complexity science to psychotherapy research. In turn, these concepts and methods can help to increase the effectiveness of therapeutic treatments. Also, having a common organizational structure may benefit science in general.