A new bachelor’s thesis at the ZHAW is the first to test whether high sensitivity shows up in the event-related potentials of the EEG. The result sits poorly with the widespread image of a permanently over-aroused nervous system – and puts a different quantity at the centre instead: vigilance.
The picture of a permanent alarm
High sensitivity – technically Sensory Processing Sensitivity (SPS) – is one of those terms that has made its way out of research and into everyday language. Roughly 20 to 30 per cent of the population show a heightened expression: they notice fine nuances, react more strongly to sensory stimuli, process experiences more deeply – and tend towards overstimulation in stimulus-rich environments.
The popular version quickly turns into a simple picture: the nervous system of highly sensitive people is permanently ramped up, on constant alert, as it were. From a neurophysiological point of view, that is a claim about vigilance – about the tonic level of wakefulness and activation a brain maintains, and about how stably it regulates that level. Vigilance is not a property of any single stimulus but the background against which all stimulus processing takes place.
This reading makes a testable prediction: anyone running permanently higher would have to respond more strongly to stimuli at the very first, purely sensory stage of processing. That is exactly what the EEG can measure. Michelle Lade did so in her bachelor’s thesis at the ZHAW – and found a more differentiated pattern.
The research question
The theoretical starting point is the predictive processing framework. Greven and colleagues (2025) explain high sensitivity within it through increased precision weighting: the brain consistently rates sensory prediction errors as particularly reliable, so even weak signals are passed upwards instead of being filtered out. From this the authors derive a clear neural prediction – higher sensitivity should show up as enhanced early ERP amplitudes.
This prediction was tested on data from the Brain and Trauma Foundation Grison, collected under the examination conditions of the Human Brain Index method. The sample comprised 435 adults between 18 and 40 years of age (mean 28.9 years, 242 women, 193 men) with pronounced ADHD symptoms. Sensitivity was assessed with a 23-item version of the Highly Sensitive Person Scale (internal consistency α = .86), brain activity with a 19-channel EEG during the visual continuous performance test (VCPT) – the potentials analysed were those elicited by the second stimulus of the NoGo condition, occipital (O1, O2) and temporal (T5, T6). The analyses used partial correlations controlling for gender.
The ADHD sample is no accident: because the spread of sensory processing characteristics is particularly wide in this population, associations can be tested with greater statistical sensitivity than in an unselected general population.
Three components, three processing stages
To understand the findings, what matters is what the three components examined represent functionally:
The P1 (in the present analysis window roughly 80–110 ms after stimulus onset) is regarded as a predominantly exogenous component. Its amplitude depends on the physical properties of the stimulus – and on the level of excitation of the sensory system. It is therefore the one of the three that lies closest to vigilance.
The N1 (roughly 130–170 ms) is modulated by attentional processes. Here stimulus features are analysed and compared with stored internal representations – top-down rather than pure bottom-up transmission.
The P2 (roughly 205–250 ms) stands for the categorisation and evaluation of incoming stimuli as well as for the regulation of task-irrelevant ones.
The genuinely interesting dividing line runs between the first and the second of these stages: first the level of arousal, then directed attention.
What was found
First, the P1 turned out smaller at higher sensitivity – occipital (r = –.09) as well as temporal (r = –.10). That is exactly the opposite direction to the theoretical prediction.
Second, the N1 turned out larger (more negative), in line with the hypothesis and consistent across both regions (r = –.11 in each case).
Third, the exploratorily examined P2 showed a smaller deflection over the temporal electrodes (r = –.11), while nothing emerged occipitally.
Fourth, the well-known gender difference in self-report replicated clearly: women scored higher on sensitivity than men (2.73 versus 2.27, d = 0.81 – a large effect). The associations between sensitivity and brain activity, however, did not differ between the genders.
The vigilance reading
The dissociation between P1 and N1 is the actual finding of this thesis – and vigilance offers the most parsimonious account of it.
If high sensitivity were a matter of a generally elevated level of activation, the vigilance-near P1 would have to grow along with it. It does not; it shrinks. What grows is the component one stage later – the one that depends on attentional control. Sensitivity thus appears here not as «more current in the system» but as a different distribution: the brain does not hold a higher level, it deploys the level it has more selectively.
The P2 finding fits the same line. A smaller temporal deflection is taken as an indication of more efficient stimulus evaluation – whoever looks more closely early on has less to re-evaluate later. Effort is shifted forward, not increased overall.
For practice this is a shift with consequences. In these data, openness to stimuli appears as a style of regulation, not a state of arousal. On this pattern, a highly sensitive person is not chronically over-activated but works with a different weighting between incoming signal and internal model. Overstimulation would then not be the consequence of a permanently excessive baseline, but the price of a very widely opened selection stage in an environment that offers too much at once.
Notable in this context is the convergence with another thesis from the foundation’s orbit: in Chanson and Assmann (2026) the vigilance parameters did not differ between ADHD with and without comorbid depression, whereas the arousal values differed clearly. In both studies, then, vigilance behaves like an axis in its own right – not like a proxy for symptom load or trait expression.
One important qualification belongs here: the present thesis did not measure vigilance directly. It is an ERP study, not an analysis of the resting EEG. The bridge to vigilance runs via the well-documented dependence of the P1 amplitude on the level of excitation of the sensory system. The reading proposed here is therefore hypothesis-generating – testable, but not yet tested.
What this means for diagnostics
Soberly considered, the associations are small: r = –.09 to –.11. Neither the N1 nor the P2 amplitude is therefore suitable as a diagnostic marker for high sensitivity; statements about individual cases are ruled out on this basis. The thesis itself says so unambiguously.
Two points nevertheless stand. First, high sensitivity shows up in measurable brain potentials at all – it is not a pure phenomenon of self-perception. That is an argument against trivialising the concept and, with equal force, against pathologising it.
Second, the gender finding is practically relevant: women report considerably more sensitivity, yet their brains show no different pattern from men’s. This suggests that the difference arises in response behaviour – in Western cultures, sensitivity carries negative connotations for men. Anyone working with sensitivity questionnaires should allow for the possibility that men systematically report less openness to stimuli than they show neurophysiologically.
Limitations
The thesis names its limitations carefully. Only adults with ADHD characteristics were examined, without a control group – the negative P1 finding could therefore be attributable, at least in part, to the P1 reduction described in ADHD rather than to sensitivity. The 19-channel setup permits no source analysis; the occipital/temporal distinction rests on scalp electrodes. The 23-item version of the HSP scale used here departs from the 27-item version common in the research literature. And the effects found lie below the threshold that the sensitivity analysis would have required for a statistical power of 80 per cent (r = .12) – replication is not optional but a precondition.
Looking ahead
The most obvious next step would be a 2×2 design comparing people with and without ADHD as well as with low and high sensitivity – only then can the contributions of the two constructs to early visual processing be cleanly separated. Equally obvious, and decisive for the interpretation proposed here: linking the ERP findings directly with vigilance parameters from the resting EEG. Only then could it be tested whether sensitivity really operates at the level of selection rather than activation.
For anyone who wants to go deeper into these methods: the Biomarker Workshop 2026 of the Brain and Trauma Foundation Grison takes up exactly these topics. Details and registration at Registration.
Many thanks to Michelle Lade for a piece of work that takes an attractive hypothesis seriously enough to let it actually fail – and wins the more interesting question from that failure.
Source: Lade, M. (2026). Sensory Processing Sensitivity: More Than Just a Feeling. A neurophysiological investigation of the associations of Sensory Processing Sensitivity using visual event-related potentials (P1, N1 and P2). Bachelor’s thesis, ZHAW Zurich University of Applied Sciences, School of Applied Psychology. Supervisor: Dr. phil. Andreas Müller.