The change that turns everything upside down

Autistic Mirror (Outside view/surrounding people, education professional)

For autistic children, the transition from nursery to school is not a simple change of place, but a fundamental change in the prediction environment. At nursery the child had built a stable internal model over months: familiar rooms, familiar faces, familiar routines. The predictive coding system, which constantly compares expectations with actual input, could work efficiently because most predictions were correct. With the start of school this model is largely unusable. New teachers, a new timetable, a new social structure, new room acoustics: every single prediction error costs processing energy, and in the first weeks such errors arise minute by minute.

The stomach ache in the morning is not pretending and not psychosomatic "tricking". Sustained predictive uncertainty keeps the cortisol level raised, and the stress axis between brain and gut runs in both directions: chronic activation reliably shows up as gastrointestinal symptoms. The child literally feels something that is real.

The withdrawal during the day and the breakdown in the afternoon are connected. At school the child activates all available resources in order to join in: it filters noise actively instead of blocking it out automatically, reads signals consciously instead of intuitively, holds back bodily regulation strategies because the context demands it. That is masking: the process of adapting one's own processing profile so that it is less visible on the outside. Masking works, but it is neurobiologically expensive. It does not run for free, but on credit. Capacity used up during the day is no longer available in the afternoon.

The breakdown after school, often called after-school restraint collapse, is exactly that: the nervous system drops its controlled stance as soon as a safe environment is reached. It looks like misbehaviour, but it is the opposite: the child has functioned all day, and at home it is safe enough to stop functioning. The secure attachment figure at home draws the collapse in rather than causing it, precisely because the child no longer has to perform there.

The withdrawal at school itself follows a different logic. When predictive energy is exhausted and new social demands arrive at the same time, the radius within which the child can still act shrinks. Withdrawal is then resource management: less social interaction means fewer unpredictable inputs that have to be processed.

What sharpens the picture further: school is sensorily more demanding than nursery. Classrooms are louder, groups larger, the change between activities is timed faster and, literally, rings. Each of these stimuli, which at nursery was still within processing capacity, can now reach the limit, because the baseline is already higher due to the general predictive uncertainty.

The child is not breaking down for a new reason. The school is doing nothing wrong, and neither is the child. What is breaking open are the costs of a system that now demands considerably more predictive effort, considerably more stimulus filtering and considerably more social adaptation at the same time than the child ever had to manage before.

Autistic Mirror (Outside view/surrounding people, education professional)

Nurseries observe the masked child. What they describe as "coping well" is often a child who has learned to suppress their actual reactions, to appear inconspicuous and to adapt to the demands of the group. That is masking: automatic autistic reactions are continuously dampened prefrontally, stimming is held back, overload is hidden. This process runs as a dual task alongside what is actually happening in the group and uses capacity that is really needed for learning and regulation. What the nursery sees is the result of this effort, not the underlying state.

This means: the knowledge a good practitioner has built up about a child is largely implicit. Without having written it down, she knows that the child needs five minutes when changing rooms, that the seat by the window is too bright, that certain textiles lead to states of emergency, that a brief announcement before tidying up prevents meltdowns. This knowledge is encoded in the relationship and in lived routine, not in a handover file. It is quiet in origin, has built up over months and is lost when the nursery period ends.

For autistic children of preschool age this loss of knowledge is particularly consequential, because their nervous system depends on predictability. Predictive coding describes how the brain constantly builds predictions about the environment and compares them with actual input. In the autistic nervous system the actual sensory value is weighted more strongly than the prediction, which means: every deviation from what is expected costs more energy than in neurotypical children. A new environment, a new key person, new routines, different noises, different smells, different rules. All of this is new at the same time at the start of school. The system has to process each of these channels actively, because no established prediction exists that could take over the work.

What is missing when the nursery only passes on that the child is coping well is exactly this: context-specific knowledge about the conditions under which the child copes. Which structures, which routines, which sensory adjustments, which ways of communicating carry this apparent inconspicuousness. Without this knowledge the receiving setting starts from zero. It first sees a child who may function well, and only begins weeks later to understand why certain situations escalate. During this transition the child absorbs the burden of the missing adjustment alone, often to the point of exhaustion.

What can help structurally is a written handover that does not contain judgements ("copes well"), but concrete descriptions of conditions: what works, with what, under which circumstances, what does not, and what the child cannot yet communicate in words. Ideally this handover is written together with the parents, because the parents see the unmasked child, the nursery the masked one.

Autistic Mirror (Outside view/surrounding people, education professional)

Even when a child knows the sequence ("after breakfast circle comes free play"), knowing the sequence does not automatically trigger the change. The mechanism behind this is called autistic inertia. The brain holds an active state stable, because the basal ganglia and the resting-state circuits connected to them are wired in such a way that moving into a new state costs disproportionately much energy. It is not blocking the change out of defiance or unwillingness, but a neurological braking process: the active state is held as long as no sufficiently strong signal arrives to set the switching process in motion.

Added to this is monotropism. Autistic attention bundles intensely on a single channel instead of spreading widely. When the child is busy with building blocks, they are really completely there, not half there and half already at the next thing. Leaving the tunnel therefore costs not just a normal transition, but a complete break of an active processing connection. That is structurally more expensive than with more widely distributed attention, and these costs do not change because the nursery routines are familiar.

That the routines are familiar still helps the child, but in a different way than often expected: it reduces fear of the unknown, not the switching costs themselves. Predictive coding describes how the brain constantly calculates predictions about what is coming and compares incoming sensory information against them. If the next situation is predictable, these comparison processes run more calmly. That creates security, but it does not resolve the inertia, because the brake does not sit in the prediction system, but in the switching system.

In the first weeks of school an external limit is added: the child's sensory budget is considerably strained by the new environment, new faces, new noises and new expectations. This continuous load is not a weakness, but the result of a stimulus filter that lets more sensory information through than in neurotypical children, because filtering out redundant stimuli before awareness runs less automatically. The fuller the budget, the more expensive every change becomes, because the capacity for switching comes from the same pot.

What helps structurally are announcements in time, not as a reminder of the sequence, but as a signal to the switching system: "In two minutes we will tidy up" gives the brain a lead time to let go of the active state slowly instead of being interrupted abruptly. The more concrete and temporally stable these transition signals are, the more predictable the switching process itself becomes.