However, this is where IIT-Gandhinagar (IIT-Gn) makes a heroic entry with work that uses the scientific script to produce a breakthrough thriller: the institute has devised a way to predict ‘audience connect’ with 72% accuracy.
IIT-Gn researchers analyzed brainwaves captured through electroencephalogram (EEG) from viewers to make sense of responses. Another study in the field of ‘neuro-cinematics’ monitored reactions to the ‘navarasas’, the nine common emotions evoked in an audience by an artistic work. The study determined that ‘raudra’ (anger) and ‘hasya’ (laughter) had the greatest provocative power among the rasas.
‘Understanding Consumer Preferences for Movie Trailers from EEG Using Machine Learning’ was recently published on the arXiv repository. The authors are Pankaj Pandey, Raunak Swarnkar, Shobhit Kakaria, and Krishna Prasad Miyapuram from the Centre for Cognitive & Brain Sciences at IIT-Gn. They took into consideration posters and trailers of 12 Hollywood movies.
The movies included “American Made”, “Battle of Sexes”, “Brad’s Status”, “The LEGO Ninjago Movie”, “Bye Bye Man”, “Murder on the Orient Express”, “Black Panther“, “Jungle Book”, and “Dunkirk”. The choices were made to cast the net across genres.
Respondents were asked to watch trailers and rate the movies based on that experience. They were also told to indicate their willingness to spend money to watch a particular film.
“Using machine learning (ML) techniques, we predicted the ‘audience connect’ with 72% accuracy. When viewers liked what they saw, a larger part of their brains was engaged, indicating interest,” said Prof Miyapuram.
“While this is a limited study, it can pave the way for an effective predictor of audience choices based on understanding consumer behaviour.”
Prof Miyapuram said that the study considered five brainwaves: delta, theta, alpha, beta, and gamma. The analysis was based on high-density EEG collected through a specially designed cap with electrodes to measure brain activity.
Such scientific endeavours are a hit with the film industry. Subhash Ghai, a well-known filmmaker, had visited the team. He told researchers that the box-office outcome is determined by factors such as storyline, cast, and budget. As for the project that focused on neural responses to the navarasas, the team factored in Indian sensibilities.
As part of a different study Dyutiman Mukhopadhyay, a postdoctoral researcher at IIT-Gn, showed viewers clips with the rasas from movies such as “Titanic”, “Modern Times”, “3 Idiots”, “Ghajini”, “Mr India”, “Braveheart”, and “Bhoot.”
The results suggested that among the negative rasas, anger captivated the audience the most, whereas laughter was the star among the positive emotions. ‘Raudra’ had the highest attention span (fixation) among all the rasas. Audience focused on eyes when negative rasas were displayed and on mouths when positive rasas played out.
Further studies are focused on eye movement patterns and emotional association with audio-visual stimuli, Prof Miyapuram said. Some of the data generated by the studies has been put into public domain for further research, he said.
