When masking hides the risk

Autistic Mirror (Outside view/Surroundings, therapist)

Standardised risk screenings measure almost exclusively what is visible: verbalised hopelessness, affective expression, agitation, crying, direct reporting of suicidal thoughts. In autistic clients, none of these indicators is reliable.

Masking is the central mechanism here. Autistic people often develop, over decades, a detailed behavioural layer that is activated in social situations: a calm voice, controlled gaze, answers that match expectations. This is not conscious deception but an automated protective mechanism that operates at the level of behavioural planning, long before a conscious decision is made. In the therapy process this means concretely: the session itself is, for many autistic clients, a social context in which masking is automatically active, even when the therapeutic setting is subjectively experienced as safe. "Switching off" masking is not a question of trust alone; it takes time, energy and often explicit permission, which standardised formats rarely give.

On top of this comes a second mechanism that structurally undermines screening: alexithymia. Emotional states are experienced but not reliably translated into language. Questions such as "How are you feeling?" or "Have you had thoughts of harming yourself?" assume that the person has internally registered what is going on and can communicate it in a socially expected form. With alexithymia, both are often missing. An autistic person in a serious crisis can answer the questions because they answer the expectation structure of the question, not their own inner state.

Autistic burnout as a background constellation considerably sharpens this problem. Burnout is a state of chronic capacity exhaustion that shows itself as a loss of functions that were previously present. It arises from sustained masking demands, constant sensory load and missing recovery. From the outside, burnout can look like stabilisation: less movement, less expression, less reactivity. Internally it is the opposite. The risk screening then rates "calm" as an indicator of safety, while calm here is exhaustion that forces the nervous system to shut down to a minimum.

The research on suicide risk in autistic people is clear and should change how screening is calibrated: autistic people have statistically markedly elevated rates of suicidal thoughts, suicide attempts and suicide. This is not a marginal difference. At the same time, they show less often the external distress signals that screening instruments measure.

What this concretely means for screening: placing less weight on affective expression and more on losses of function. Questions such as "Are there things you can no longer manage lately that used to work?" or "How is your energy in everyday life?" capture autistic burnout more directly than emotional self-report. At the same time, it helps to name masking actively in the session: "I know it can be exhausting to appear 'in shape' here. You don't have to." This is not a conversation script but a framing that counters the automatic activation of masking.

The crisis plan, if one exists, should have been worked out explicitly with the autistic person, taking into account the sensory barriers that a clinic or a psychiatric emergency department represents for autistic people. Without this step, a crisis plan that provides for hospital admission in an emergency may be less effective than assumed.

Autistic Mirror (Outside view/Surroundings, therapist)

Body signals such as tension, exhaustion or overwhelm arise in the insula, a cortical region that continuously integrates internal states and passes them on to awareness. In many autistic people this signal arrives at a lower resolution, either dampened, delayed or in thresholds rather than as a gradient. This means: high strain is not necessarily felt as "high", because the body signal itself does not arrive in proportion to the actual state. That is interoception. The brain registers the state, but the internal reporting system does not deliver a reliable volume.

Alexithymia comes in one step later. Even when a body signal arrives, the automatic translation step from body state to feeling word is missing. "I sense something" does not reliably become "I am exhausted" or "I am overwhelmed". Anyone asked on a scale of 1 to 10 needs exactly this translation step. The answer can therefore be low, not because the person is downplaying the situation, but because access to their own inner state is simply not available at the moment of the question.

Both mechanisms work together here, but as separate levels. Interoception is receiving the signal. Alexithymia is assigning the signal to a nameable state. An autistic person can have difficulties with both, with one of them, or the connection fluctuates situationally. Under high strain itself, interoceptive resolution is often reduced even further, because cognitive resources are tied up elsewhere. That means: precisely when the state is most severe, it is hardest to access.

Clinically relevant is that the number given should not be treated as unreliable self-report, but as what it is: an honest statement based on incomplete inner data. Scale questions assume that the person asked can make a reliable inside-outside comparison. For many autistic clients this is not a stable mechanism.

More productive than "How strained are you on a scale?" can be body-near or behaviour-based approaches: What is harder today than usual? Are there things that normally work that don't work today? These questions bypass the interoceptive translation step and instead tap observable differences that are accessible even without the capacity to name feelings.

Autistic Mirror (Outside view/Surroundings, therapist)

Gradual withdrawal and a flat tone of voice in autistic people often arise from a different origin than what neurotypical communication frameworks read into such signals. Neurotypical diagnostics interpret social withdrawal and flattened affect as an indication of depressive or suicidal states, because in neurotypical processing they are indeed often precursors or accompanying features of these. In autistic people, the same outward signs are often an expression of autistic burnout, a state that is mechanistically different.

Autistic burnout arises from cumulative masking load. Masking means that autistic processing and communication styles have to be permanently translated into neurotypically readable forms, which binds prefrontal control that is not available for other processes. At the same time, sensory filtering runs without sufficient pre-attentive habituation, that is, consciously and glucose-expensively rather than automatically dampened. This constant load accumulates over weeks and months. The withdrawal is then not a signal of hopelessness but a regulation strategy: fewer stimuli, fewer masking demands, less capacity loss per unit of time. The flat tone follows the same logic. Expressing emotion outwardly is itself an effort that costs capacity. When the budget is exhausted, expression control is switched off first, not because there is no emotion, but because the resource to transmit it is missing.

The distinguishing feature from acute risk lies not in outward appearance but in the course over time and in the inner structure. Autistic burnout shows itself as a loss of abilities over time: things that worked as a matter of course three or six months ago are currently out of reach. Speaking, opening letters, keeping appointments can be affected, not as the result of a decision, but because the neurobiological capacity is simply no longer there. The pattern is gradual, directional and can often be traced back to a period of beginning overload.

Acute risk follows a different pattern. The empirically established risk markers for suicidality in autistic people are cumulative camouflaging, burnout as a precondition, late diagnosis without support, persistent involuntary loneliness and the experience that needs remain permanently unmet, not autistic traits as such. The chain runs through sustained masking, its cortisol load and the resulting exhaustion. Acute risk is therefore often not the first visible step, but a point far into an already longer-running spiral of strain. Withdrawal in burnout can be a protective mechanism, but it can also be a context in which risk arises, when exhaustion has progressed so far that no recovery seems possible and at the same time there is no access to support.

The double empathy problem is clinically relevant here. Autistic people often communicate strain differently from what neurotypical communication systems expect: more directly, more factually, without the emotional framing that neurotypical clinicians take seriously as a signal. At the same time, implicit, masked signals of strain often remain invisible, because the translation work is missing in both directions. The double empathy problem describes that this gap in understanding is mutual: the autistic person communicates in a way that the neurotypical counterpart does not read as a distress signal, and the neurotypical counterpart interprets in a way that does not reach the autistic person. In therapeutic and diagnostic contexts this concretely means that standard screening questions at the emotional meta-level often do not reflect what is actually present.

A differentiated assessment therefore needs questions that start at the level of ability: What was possible six months ago that is no longer possible now? How long has this course been running? Are there current thoughts of seeing no way out that are named explicitly, even if the tone stays flat? A flat tone when answering this question is not a contraindication; it corresponds to what is available in the burnout state.