P53. Optimizing ERP detection in disorders of consciousness by taking into account pre-stimulus brain states

Abstract

Objective Two conditions can be summarized as disorders of consciousness (DOC) - the unresponsive wakefulness syndrome (UWS) and the minimal consciousness state (MCS). The first is defined by the total absence of awareness while in the latter, patients show some marginal signs of awareness but in a somewhat inconsistent way. At present, the extent to which a patients is capable of processing environmental information is still unclear. To address this questions, many researchers have used event related potentials (ERPs), as means to assess cognitive capabilities in the absence of overt behaviour. However, the extent to which studies find ERPs in DOC patients varies greatly. For example, the P300 has been demonstrated in 100% of DOC patients in a study by Schnakers et al. (2009; 22 DOC patients) but only in 18% in a study of Witzke et al. (1996; 12 of 66 DOC patients). This wide range might be a result of the fact, that there is no gold-standard in analyzing DOC patients ERPs. However, most DOC ERP studies are based on classical averaging methods which need several artifact-free epochs to average, but do not take into account the pre-stimuli brain state. This makes a great many repetitions of stimuli a necessity. However, reports from recovered DOC patients indicate, that patients rapidly drift in an out of awareness. So, it is possible, that in typical averaging methods, the few actual P300 responses get averaged away because of lots of ‘missed stimuli’ that the brain did not register. At the same time, results from basic research indicate that the likelihood of stimulus registration is related to pre-stimulus brain activity, in particular to alpha band power. We therefore use a method that help to predict that a stimuli might be missed, so this trials could be excluded from further analysis. Method In the undamaged brain, activity in the alpha band is related to inattentiveness or inhibition of external stimuli. Therefore, we reanalyzed ERP date of 17 DOC patients in such a way that we first determined pre-stimulus alpha power and then sorted trials accordingly, excluding those with dominant alpha activity from further analysis. We tested the patients’ data for a significant P300 reaction by means of standard averaging methods (every artifact free trial) in comparison with the analysis of only those trials with the least amount of alpha band power. Results With the standard averaging methods we found a significant P300 in only a few patients. However, using the alpha sorted trials, we were able to demonstrate significant P300 reactions in patients, where the standard method did not detect an ERP. Implication With standard averaging methods the amount of patients able to respond with cognitive ERPs might be underestimated. Further research assessing the cognitive abilities of DOC patients should aim to deliver stimuli either at the optimal moment, or try to separate post hoc optimal phases from those, where patients were unable to respond.