ausblenden:
Schlagwörter:
-
Zusammenfassung:
The ability to track the statistics of our surroundings is a key computational challenge. Dayan
and Yu [1] proposed that the brain monitors for unexpected uncertainty – events which
deviate substantially from model predictions, indicating model failure. Norepinephrine (NE) is
thought to play a key role in this process by serving as an interrupt signal, initiating modelupdating.
To determine whether NE routinely reports the statistical structure of our surroundings, we
used rapid tone-pip sequences that contained perceptually salient pattern-changes
associated with abrupt structural violations vs. emergence of regular structure (stimulus
example: https://bit.ly/2KzVXLu).
Participants were instructed to detect short silent gaps within the sequences. This ensured
broad attention to the auditory stimuli but without requiring active tracking of the transitions.
We found that even though both transition directions (regular-to-random and random-toregular)
are clearly detectable behaviourally and both evoke strong MEG [2] and EEG [3]
responses in naïve distracted listeners, only abrupt structural violations (regular-to-random)
evoked pupil dilation. This pattern of results demonstrates that, when pattern transitions are
not behaviourally relevant, NE tracks unexpected uncertainty on rapid time scales relevant to
sensory signals.
In a following experiment, we sought to understand how pupil responses are affected by
behavioural relevance. We asked participants to monitor for and report both types of
transitions. Marked differences in pupil dynamics were observed. Most notably, active
monitoring gave rise to a pupil dilation response to the emergence of regularity. Importantly,
this response was not strongly linked to the execution of a motor command as response time
accounted for relatively little variance in various pupil diameter metrics (e.g. change in pupil
diameter, pupil diameter derivative, etc.) and this effect was preserved in a delayed response
version. These behaviour-related changes in the pupil diameter suggest that behavioural
relevance may alter the boundary between different types of uncertainty (e.g.
expected/unexpected), resulting in a threshold change for model reset.