Nijhawan reviews the need for theoretical consideration of delays brought about by differential neural transmission time. Such a need is unequivocal in some situations, and the flash-lag effect, which Nijhawan has championed for some time, seems reasonably accounted for in such terms.

However, not all neural delays need be compensated for. Instead, there is often wide tolerance for integrating diverse continuous signals. As an example, consider first the asynchronous presentation of auditory speech and its visual signal. What are the tolerances in the perceiver for detecting that anything is amiss? As it turns out, the window is fairly large and asymmetrical. The auditory delays referenced to the visual signal can be between −45 msec (an auditory lead) and +200 msec (an auditory lag; Grant et al. Reference Grant, van Wassenhove and Poeppel2004). Rewritten, this is ~80±120 msec. The 80 msec value can be taken as a difference in the delay between the auditory and visual systems. It might need to be “compensated” for, and it is consistent with Nijhawan's report. However, the tolerance window of ±120 msec from that mean value suggests that no compensation is necessary. Instead, information in both signals is temporally distributed and, arriving differentially from auditory and visual modalities, is seamlessly integrated within a fairly wide temporal window.

This first example, however, is not about perception and action, with which Nijhawan is most concerned. Therefore, consider a second case – the disfluency effects of delayed auditory feedback while speaking. Stuart et al. (Reference Stuart, Kalinowski, Rastatter and Lynch2002) reported no performance differences among 0, 25, 50 msec delays, but there were many disfluencies at 200 msec delays. In other words, the coupling of auditory perception and vocal action is no more affected by a 50 msec delay than by listening to oneself normally. This too suggests tolerance, not compensation.

Of course, this second line of evidence concerns audition, which is not the modality of Nijhawan's focus. Therefore, let us consider, third, the results from the literature on delayed visual feedback. In a functional magnetic resonance imaging (fMRI) study, Farrer et al (2007) showed that subjects had no differential awareness of a 0 versus 100 msec visual delay in watching themselves pluck pegs from a board, and their performance did not suffer either. Moreover, the angular gyrus, which seems to register awareness of such asynchrony, was not substantially engaged until delays of 150 msec and longer. Again, this suggests tolerance for décalage in signals, this time visual and motor.

Nijhawan's primary interest concerns the prediction of the visual location of a moving object in space. The most critical everyday venues for such needs are in hunting and in ball games. Among the latter – whether baseball, cricket, tennis, or ping pong – one can rightly argue for a required accuracy of ~5 msec when hitting a ball with appropriate control (Bootsma & van Wieringen Reference Bootsma and van Wieringen1990; Land & McLeod Reference Land and McLeod2000; Lobjois et al. Reference Lobjois, Benguigui and Bertsch2006; Tresilian Reference Tresilian1993; Watts & Bahill Reference Watts and Bahill1990). Nijhawan's point is that the visual system – retina to V1 and beyond – is slow by comparison.

Nonetheless, in all of these sports, the best participants are involved in highly stereotyped, highly practiced situations. Moreover, the relevant behavior of the opponent and the flight of the ball are distributed over as much as a second and often more. In baseball, those deemed excellent hitters – that is, they can accurately hit the ball 3 out of 10 times against the best pitchers – earn well more than US$1000 per swing of their bat.

Skill accrued over many years – and a background of appropriate genetics – is required to perform at this level. Moreover, there is ample evidence that hitters cannot follow the ball near the critical time when they need to hit it (Land & McLeod Reference Land and McLeod2000; Watts & Bahill Reference Watts and Bahill1990), and visual acuity in the periphery, where the ball arrives when it is hit, is extremely poor. Hitting skills are not due to higher cognition. Indeed, thinking often interferes with performance (e.g., Beilock et al. Reference Beilock, Bertenthal, McCoy and Carr2004). But such performance is also not due to the kind of neural compensation that Nijhawan addresses. Instead, it is the result of a skill-driven refinement of perception and action in the context of the information distributed in time.

A kind of compensation does occur in a number of situations. In archery, and in trap and skeet shooting, the sportsperson must aim to lead the target that is being shot at. Yet it seems unlikely that one should look for pure sensory or motor accounts in such feats. Instead, and again, it is the skill of the individual calibrated over repeated practice that allows targets to be hit.

None of this is to denigrate Nijhawan's attempt at laying out the needs for a coherent theory of dynamic neural organization. My purpose is simply to state that computational tolerance for the time-staggered arrival of integrable neural information, and skill acquired over many long bouts of practice, are likely to be more important than is compensation.