E-MOVIE - Experimental MOVies for Induction of Emotions in neuroscience: An innovative film database with normative data and sex differences

Abstract

The need for a validated set of emotional clips to elicit emotions in more ecological experiments is increasing. Here we present the validation of a new database of emotional films, named E-MOVIE, which includes, in this first validation phase, 39 excerpts arranged in six categories, three negative (Fear, Sadness and Compassion), two positive (Erotic and Scenery) and a Neutral category. Notably, Compassion and Scenery are new in the field as they were not included in other databases. The clips in E-MOVIE are characterized by homogenous durations of approximately two minutes, which make them suitable for psychophysiological research. In order to study the affective profile prompted by each category 174 participants (112 women) rated the movies on multiple dimensions, namely valence and arousal, intensity and discreteness of the induction of one of the six basic emotions and, finally, intensity of the experience of the emotional states defined by a series of emotional adjectives. Erotic clips were effective in the elicitation of a positive emotional state, characterized by high levels of arousal and excitement. On the other hand, Fear clips (selected without blood to avoid disgust reaction) prompted an affect characterized by high arousal, low valence and high levels of reported fear and anxiety. Women reported greater unpleasantness, distress, anxiety and jittery than men to the three negative categories. Compassion clips, characterized by the depiction of crying characters, were able to induce an affective state dominated by sadness and feeling touched, consistent with an empathic reaction to emotional sufferance. Sadness clips, instead, elicited an affective state characterized by sadness together with distress and angst. We also demonstrated that clips depicting natural environments (i.e. Scenery) prompted in the viewer a surprised, inspired affective state, characterized by high valence and arousal (especially in males), a result which suggests that their past categorization as neutral stimuli was inaccurate and problematic.

Spectrally resolved EEG intersubject correlation reveals distinct cortical oscillatory patterns during free‐viewing of affective scenes

“…Distribution of each emotional category ratings on the valence/arousal space from E‐MOVIE (Maffei & Angrilli, 2019a) normative data. Dots represent mean valence and arousal ratings.…”

“…The stimuli were 18 movie clips selected from E-MOVIE (Maffei & Angrilli, 2019a), arranged in six emotional categories, with three clips per category. The categories were Erotic (heterosexual couples engaged in a sexual act), Scenery (stunning views of different natural landscapes), Fear (thrilling scenes of anticipated threat), Sadness (scenes featuring themes of desperation and helplessness), Compassion (scenes of loss, grief, and bereavement), and Neutral (scenes drawn from urban documentaries).…”

“…Fear, Sadness, and Compassion were instead selected to elicit unpleasant and aversive responses, with both high arousal (Fear) and low arousal (Sadness and Compassion). In the E-MOVIE validation (Maffei & Angrilli, 2019a), each clip was rated by 174 participants on valence and arousal with the Self-Assessment Manikin (Bradley & Lang, 1994), and on how the emotions experienced during each clip fit a series of emotional adjectives to provide a comprehensive assessment of the affective state they elicited. 1 The distribution of each film score on valence and arousal is shown in Figure 1 The clips were edited, using Adobe Premiere CS5, into a single continuous clip with a resolution of 1,280 × 720 pixels, in which the clips were pseudo-randomized, to avoid the consecutive presentation of two clips of the same category.…”

“…For this study, stimuli were selected from E‐MOVIE (Maffei & Angrilli, 2019a), a recently published collection of short film clips with standardized duration, explicitly designed to be used in psychophysiological and neuroimaging experiments. These clips are designed to probe into several affective states to study reactions at both subjective (Maffei, Vencato, & Angrilli, 2015) and physiological level (Maffei & Angrilli, 2019b; Maffei, Spironelli, & Angrilli, 2019).…”

Toward Emotion Recognition From Physiological Signals in the Wild: Approaching the Methodological Issues in Real-Life Data Collection

“…The examples are the LIRIS-ACCEDE dataset (Baveye et al, 2015 ) containing continuous annotations of arousal and valence, and FilmStim (Schaefer et al, 2010 ) annotated with 6 discrete labels. Other more specific datesets include “the emotional film database for Asian culture” (Deng et al, 2017 ), which was validated through physiological measurements, or the E-MOVIE dataset (Maffei and Angrilli, 2019 ) composed of the movie extracts of fixed duration making them suitable for psycho-physiological research. Interestingly, the latter is annotated not only with emotion dimensions but also with discrete emotion labels.…”