Recovery from catatonia follows a hierarchical precision order

The human brain is constantly making predictions about the world, from the texture of a cup in your hand to the rhythm of a conversation. To navigate this flow, it must decide how much weight to give to new information versus its existing expectations. In the field of computational psychiatry, scientists use a framework called the free-energy principle to describe this process. They propose that the brain assigns a level of confidence, or "precision," to different types of signals: the raw data coming from our senses, the plans we make to move our bodies, our internal drive to achieve goals, and our understanding of how stable or chaotic our environment is. When these confidence levels are misaligned, the brain's ability to update its model of reality breaks down, leading to confusion and a loss of control over one's own actions. This theoretical lens offers a way to look beyond the surface symptoms of mental illness and ask what specific part of the brain's prediction machinery has gone wrong.

One of the most severe manifestations of this breakdown is catatonia, a state where a person may become completely still, unresponsive, or locked in a rigid posture, or conversely, become agitated and unable to stop moving. For decades, doctors have treated this condition with a standard two-step approach: first, they try to calm the patient with benzodiazepines, a type of sedative; if that fails, they often turn to electroconvulsive therapy. While these treatments frequently work, the process has been largely empirical. Clinicians have lacked a clear rule for what to do when the first step doesn't fully restore a patient's ability to function, or why some patients seem to improve physically but remain mentally fragmented. A new study from a team of researchers in Japan and Sweden suggests that recovery from catatonia is not a random collection of improvements, but a strict, hierarchical sequence. By analyzing the medical records of twenty-five patients who recovered without electroconvulsive therapy, the researchers found that the brain must repair its functions in a specific order, starting with basic sensory stability and moving up to complex, goal-directed behavior.

The researchers examined twenty-five consecutive inpatients diagnosed with catatonia at a psychiatric hospital that does not use electroconvulsive therapy. These patients had a range of underlying conditions, mostly involving schizophrenia or mood disorders, and they were treated with standard medications chosen by their doctors based on symptoms. The team did not intervene with a new protocol; instead, they looked back at the detailed notes taken during the patients' hospital stays. They applied a specific computational model to these notes, asking whether the chaotic mix of symptoms and the gradual return to health could be mapped onto five distinct levels of brain function. They defined these levels as sensory precision (how the brain handles raw sensory input), policy precision (the ability to select and execute actions), motivational precision (the drive to start doing things), fast volatility (moment-to-moment stability), and slow volatility (long-term consistency).

What they discovered was a rigid pattern. In every patient for whom the data was evaluable, the improvements happened in the exact same order. The first sign of recovery was always the calming of the body's raw, sensory reactivity. Patients stopped being overwhelmed by autonomic instability, such as rapid heart rates or unprovoked agitation, and began to rest. Only after this sensory storm settled did the next stage appear: the ability to tolerate structured care. Patients could finally sit still for feeding, allow nurses to help them wash, or walk with assistance without their behavior collapsing. This was not yet a return of will; it was simply the capacity to execute an action once it was guided by someone else.

The third stage, which followed the stabilization of action, was the return of motivation. Once patients could reliably perform tasks when guided, they began to start things on their own. They would reach for food without being asked, cooperate with oral care without prompting, or speak briefly without being led. The researchers noted that no patient ever showed this kind of spontaneous, goal-directed drive before they had first regained the ability to execute organized actions. Only after this internal drive returned did the final two stages unfold. The patients' speech and behavior, which had been fragmented and shifting from moment to moment, began to stabilize. Finally, the last piece of the puzzle fell into place: the restoration of long-term coherence. Patients regained the ability to maintain a consistent narrative and engage with the world in a stable way across different contexts and times of day.

This sequence held true for all twenty-five patients where the order could be determined, with no single case showing a reversal of the order. (Note: For three patients treated with antipsychotics from the start, the "fast volatility" domain could not be evaluated, so the sequence was not tested for them in that specific step). The average Bush–Francis Catatonia Rating Scale score, a measure of symptom severity, dropped dramatically from 26.3 at the peak of the illness to just 2.0 after recovery. The study suggests that the brain cannot skip steps in this repair process. A patient cannot regain the will to act if they are still overwhelmed by raw sensory noise, nor can they maintain a coherent story about their life if their moment-to-moment actions are still unstable. The researchers observed that simply suppressing symptoms with medication often failed to produce this full recovery; many patients were sedated and quiet but remained incoherent and unable to function. True recovery, they found, was defined by the return of organized action, which then unlocked the higher levels of function.

The team describes this process as a "Saito Loop," a self-reinforcing cycle where the stabilization of one level provides the necessary evidence for the next level to lock into place. It is not a simple linear list of symptoms disappearing, but a dependency chain where the brain must secure the foundation before it can build the roof. The study observed that in many cases, the number of medications was reduced as the patients improved, suggesting that the brain was reorganizing itself rather than just being chemically subdued, though the retrospective design precludes definitive causal claims about medication driving the recovery. The findings also challenge the notion that recovery is a chaotic or idiosyncratic event; instead, it appears to be a constrained progression that follows the same rules for everyone, regardless of their specific diagnosis or the mix of drugs they received.

While the study is based on a retrospective review of records and cannot prove the underlying mechanism with absolute certainty, the consistency of the pattern across twenty-five different cases is striking. The researchers acknowledge that their work is a hypothesis-generating study, meaning it proposes a model that needs to be tested in future, prospective trials. They note that their findings apply specifically to patients who eventually recovered, and the model does not yet explain what happens in cases where recovery fails. However, the data strongly suggests that the brain's path out of catatonia is not a random walk but a structured climb. By understanding that the brain must first stabilize its sensory input, then its ability to act, and finally its drive and coherence, clinicians may be able to better recognize where a patient is stuck in the process. This could help explain why some treatments work while others fail, and why a patient might appear calm but remain functionally broken. The study offers a new way to see the disorder, not as a collection of broken parts, but as a system that must be repaired in a specific, logical order.