On the Cognitive principles underlying operational biases in the temporal domain – CoPrOBTeD

We use a combination of classical cognitive experiments and psychophysical experiments in healthy adult participants.

1. We tested the sensory integration hypothesis by comparing filled and empty intervals in a temporal arithmetic task. Addition outcomes were consistently overestimated relative to baseline in both cases, and the bias was even stronger for emptyintervals. These results rule out a sensory accumulation account and support a cognitive origin of the TME.

2. We examined whether the TME could be explained by timing principles such as central tendency, range effects, and scalar variability. We replicated the TME across low and high stimulus ranges, with its magnitude increasing as a function of range. Although general timing principles influenced estimates, they could not fully account for the bias, highlighting its specificity as a cognitive effect.

3. We investigated the functional locus of the TME using a comparison paradigm to disentangle perceptual and motor components. We observed a reversed pattern, subtraction was overestimated and addition underestimated. This reversal suggests that the manifestation of the TME may depends on attentional allocation and methodological demands.

4. We compared the TME with the OME in a preliminary study. Both effects were observed, but they did not correlate, and their profiles diverged: the OME was primarily driven by subtraction, whereas the TME was predominantly driven by addition. This asymmetry indicates that despite superficial similarities, the two effects may be underpinned by distinct mechanisms

Our results establish the TME as a cognitive bias that does not emerge from the accumulation of sensory energy or the general principles governing time perception. This constricts the hypothesis space for further pinpointing the cognitive mechanisms underlying this bias. It remains unclear, for example, whether or not the TME can be considered as a member of the momentum effects. In particular, it has been hypothesized that a common set of cognitive mechanisms govern different momentum effects (representational momentum, attentional momentum, operational momentum). While we reveal certain parallels, our results demonstrate striking differences between temporal momentum and operational momentum (e.g. while the OM is driven by subtraction, the TME is drive by addition problems). Furure studies need to explore these effects in the same participants to systematically explore the within-subject correlation of these effects.

A surprising result was to see the reverse TME in comparison paradigms. While this implies a potential locus of the TME at the motor stage, it may also reflect a pacemaker rate that is not constant but varies with interval length: it is higher for shorter durations, providing greater precision, and decreases as durations become longer. In combination with an attentional gating mechanism, this may explain this finding. The model assumes an attentional modulation of the gate that controls the accumulation of pacemaker pulses by the accumulator instance: when attention is diverted away from temporal processing, fewer pulses enter the counter, producing underestimation. In our temporal arithmetic task, cognitive demands differed substantially from classical interval comparison tasks. Indeed, participants had to combine two intervals to internally create a third one, which was then judged against the comparison interval. Compared to reproduction paradigms, this increased cognitive load and required attention to four intervals simultaneously. One possible explanation for the reversed TME observed in Experiments 1 and 2, is that attentional resources decreased over the course of a given trial and were minimal for the second operand that did not benefit from primacy or recency effects. Specifically, attention may have been fully engaged on the first operand, but as the operation sign appeared, focus shifted towards performing the computation. According to this hypothesis, these attentional fluctuations could cause the second operand to be perceived as shorter than its actual duration. If the second operand is underestimated as a result, addition would yield overall underestimation (fewer pulses added), while subtraction would produce overestimation (fewer pulses removed). This hypothesis needs to be scrutinized in a set of future experiments.