Write a comment

This review resulted from the graduate-level course "How to Read and Evaluate Scientific Papers and Preprints" from the University of São Paulo, which aimed to provide students with the opportunity to review scientific articles, develop critical and constructive discussions on the endless frontiers of knowledge, and understand the peer review process.

The study "Pain reflects the informational value of nociceptive inputs" by Michel-Pierre Coll et al. examines the potentiality for the central nervous system to optimize pain’s learning function by being modulated by pain perception. Moreover, the author's hypothesized that nociceptive processing could be augmented when pain is surprising tofacilitate predictive learning about pain.

To see if this occurs, the authors record nociceptive flexion reflex amplitude combined with electroencephalography to link nociceptive flexion reflex modulation with the influx of ascending nociceptive signals to the cortex.

On the electroencephalography they use two indices of expected pain. The first is the late positive potential (peaks around 400-600 ms after the pain-predictive cue). The second is a decrease of oscillatory power in the alpha and beta bands.

This seems to be a very powerful take on how pain learning is associated with the expectation of pain. And the combinations of different methods of recording the perception of pain is a novelty in this area, combining the recording of nociceptive flexion reflex modulation amplitude with electroencephalography to link the modulation with the influx of ascending nociceptive information to the cortex. That way the authors can distinguish between the nociceptive responses and perception of pain in the participants and that the predictions errors between anticipated predictions and actual outcomes are associated with the nociceptive flexion reflex.

The procedure of the experiment is very well delineated and explained in the text.

The statistical analysis is very well explained and appropriate for the problem in hand. Then again, the authors did not specify why they use the three computational models that they choose to use. Although those three computational models are frequently used in studies about uncertainty in perception takes in volatile environments, the justification fortheir uses are intrinsically different. The justification for the use of a computational model must be a priori unless the study is about the models themselves. All that said, the interpretation for the basis why the Rescorla-Wagner model provided a better fit for the data is very good. Maintaining the environmental volatility relatively high and constant throughout the task due to the regular introduction of new cues is a very good way of simulating actual environmental conditions.

Also, they did not say why they used thirty-five participants. Is always recommended to do a pilot to calculate effect and sample size or to justify this quantity from the literature or previous works from the group, published or not. It is argued that this is good enough to reach meaningful conclusions in this specific case.

A limitation of this study is the dependence on the connection of the electroencephalography markers used for pain processing to aversive prediction errors that still don't have strong evidence in the literature. The authors make that clear in the text. Meanwhile, combining electroencephalography with functional magnetic resonance imaging or functional near-infrared spectroscopy may give more insights and give more evidence on pain processing to aversive prediction.

The results of prediction errors being associated with increased pain perception and physiological pain responses and also with increased anticipatory cortical responses shows that this study is well done and very promising reliability and validity.

Competing interests

The author declares that they have no competing interests.