The Price of Breaking the Tyranny of Film: The cognitive demand of top-down processes

The Role of Event Understanding in Guiding Attentional Selection in Real World Scenes: The Scene Perception & Event Comprehension Theory (SPECT)

“…Indeed, participants' in the map task condition, who saw the film clip from the beginning (i.e., when the bomb was placed in the car) were far less likely to make a prediction about the bomb exploding than those who simply watched for comprehension (i.e., 13% vs. 50%), strongly suggesting that the map task interfered with comprehending the narrative. A follow-up study (Simonson et al, 2020, June) replicated and extended this result to a set of eight film clips, and two viewing tasks: watching the clip in preparation to do 1) the map task, versus 2) a comic drawing task (i.e., watch the clip in preparation to convey the film clip narrative in a 4-panel comic). Attentional selection differed between the tasks, with only the map-task producing fixation patterns away from the main characters (entities).…”

“…Importantly, we have also shown that executive attentional control can volitionally override the impact of stimulus saliency of a dynamic film stimulus when the task is at odds with comprehending the narrative (Hutson et al, 2017;Simonson et al, 2021). In Simonson et.…”

“…Overall, task-driven top-down influences seem to be at the top of the hierarchy; whether in still images (Einhauser et al, 2008;Henderson et al, 2007;Underwood & Foulsham, 2006) or in filmed events if their composition provides the time and space for scene exploration (Hutson et al, 2017;Simonson et al, 2021;Smith & Mital, 2013). Dynamic scenes that are composed to focus attention on key features and limit individual exploration, such as movies or video clips, have been shown to allow less expression of the event model (Hutson et al, 2017;Loschky et al, 2015) and task influences (Taya et al, 2012).…”

“…Thus, in the absence of a task, both saliency (Carmi & Itti, 2006;Mital et al, 2010) and knowledge-driven selection (Hutson et al, 2022;Smith & Martin-Portugues Santacreu, 2017;Valuch et al, 2017) can rapidly influence the first saccade in a scene, with semantics having more of an influence after roughly 2,000ms (Carmi & Itti, 2006). But in the long run, task-driven attentional selection tends to overcome the influences of saliency (Einhauser et al, 2008;Henderson et al, 2007;Hutson et al, 2017;Simonson et al, 2021;Smith & Mital, 2013;Underwood & Foulsham, 2006).…”