An Invited Comment on 'Abusive Head Trauma through Shaking: Examination of the Perpetrators According to Dating of the Traumatic Events' by Laurent-Vannier et al. (2021)T he article by Laurent-Vannier et al. (2021) in this issue of Child Abuse Review is a compelling study with new and useful findings. Nonetheless, the article raises questions regarding the certainty of conclusions on two issues: determination of the mechanism of inflicted head injury and clinical dating to identify perpetrators of abusive head trauma (AHT). This commentary will discuss these two issues and why they are important.In their case series, Laurent-Vannier et al. (2021) indicate that shaking was the main and constant mechanism of injury. While some of their cases include confessions of shaking alone, most do not have confessions. Among those cases where the perpetrator did not confess, the authors present no clinical findings that indicate the presence or absence of impact. Therefore, it is hard to conclude whether impact played a significant, or even perhaps primary, role in at least some of the injuries. Accidental head injuries, resulting mostly from impact, and inflicted head injuries, resulting from either shaking or impact, or a combination of the two, overlap considerably (Christian, 2007;Kemp et al., 2011). Confessions (Edwards et al., 2020 and eyewitness accounts can help determine the mechanism of injury, but they can sometimes be false. Therefore, the determination of the mechanism of an inflicted head injury is often difficult. This difficulty in specifying the mechanism is in part why the American Academy of Pediatrics has recommended the use of the term 'abusive head trauma' that does not imply a single injury mechanism (Christian and Block, 2009). Laurent-Vannier et al. (2021) use clinical dating to identify perpetrators in their case series. The onset of symptoms from AHT very likely occurs immediately or shortly after the abusive event. Therefore, the time of onset of even mild symptoms may be the best clue for dating the incident (Adamsbaum et al., 2014). Presumably, if the time of the abusive event is known, then who was present with the victim might be determined and the