Replicability and impact of statistics in the detection of neural responses of consciousness: Table 1

Sir,We read with great interest the correspondence between Tzovara and colleagues and Naccache et al. regarding the neural detection of sound sequences in the absence of consciousness (Naccache et al. , 2015; Tzovara et al. , 2015 a , b ). We believe that this discussion raises important questions on the way to use event-related potentials (ERPs) for predicting the recovery of comatose patients. Two aspects of the correspondence particularly caught our attention: (i) the difficulty in obtaining a perfect neural detection in the control group of fully conscious subjects; and (ii) the variability of the statistical analyses used to detect neural signs of awareness.First, in their replication of the elegant local global paradigm conceived to evaluate consciousness, Tzovara et al. could not obtain a significant conscious effect in all aware control subjects, which contradicts the perfect performance obtained in the original studies (Bekinschtein et al. , 2009; Faugeras et al. , 2011, 2012; King et al. , 2013). Such a decrease in performance has also been observed in the replication of other experimental designs used to detect signs of awareness (Owen et al. , 2006; Monti et al. , 2010) both by others (Comte et al. , 2015; Gabriel et al. , 2015) and the same research team (Fernandez-Espejo et al. , 2014). Replication studies are vital for clinical and scientific progress and the slightest change in the design or in the population tested can dramatically alter the detection of awareness, as in the replication performed by Tzovara and colleagues. The experience of the clinical team is also of major importance. For example, in the local …