Assessing the Therapeutic Potential of Targeted Attentional Bias Modification for Insomnia Using Smartphone Delivery

Creation of experimental word stimuli Research assessing the presence of attentional bias in insomnia has consistently employed paired word stimuli that contain a threatening member, which communicates a meaning relevant to the concerns of those with sleep disturbance, and an emotionally neutral stimulus member (see [1]). In line with this, we included 48 threatening sleeprelated words characteristic of pre-sleep worry thoughts, paired with neutral words matched for length and spoken frequency [2] for the purpose of modification and assessment of attentional bias. The threat words were selected from an initial pool of 96 candidate words. These words were generated on the basis of three factors identified by Wicklow and Espie [3] as core themes relevant to presleep worry consistently reported by individuals with insomnia: active problem solving (e.g. thoughts about the negative consequences of poor sleep), present-state monitoring (e.g. thoughts about bodily functions), and reactivity to external stimuli (e.g. thoughts about environmental factors disturbing sleep). All 96 candidate words were rated by 12 independent judges using a 7-point scale on emotional valence, from 1 (Very negative) to 7 (Very positive) and relatedness to sleep concerns, from 1 (Not at all) to 7 (Extremely). The final selected threat words were rated to be strongest in negative valence (M = 2.35, SD = 0.78) and most relevant to sleep concerns (M = 4.67, SD = 0.93). The word pairs were randomised and divided into two separate stimulus subsets (Subset A and Subset B). For half the participants in each condition Subset A was used in the training/control component and Subset B was used in the assessment component of the task, while the remaining participants received the reverse allocation. The use of different stimuli in the training and assessment components ensures that any change in attentional bias can be attributed to the class of stimulus and not to specific stimuli themselves. These stimuli are provided in Supplementary Table 1.