Senal Dharmasiri

The cinema has always been suspended between two poles: on one side, the aesthetic and the subjective, the ineffable quality of moving images that seize the imagination; on the other, the technical and the measurable, the way editing, sound design, and pacing shape perception in quantifiable ways. My proposal for a decision-support tool (DST) for filmmakers, built from the fusion of neuroscientific insights and computational analysis of audiovisual features, tries to inhabit this contested middle ground. It is an experiment in taking seriously both the fragility of human subjectivity and the promise of data-driven objectivity. But the very attempt to bridge these worlds requires us to embrace compromises, limitations, and philosophical humility.

The central question is deceptively simple: can we turn neurocinematic measures and audiovisual analytics into something actionable for editors in the post-production process? Can we translate EEG synchrony, eye-tracking dispersion, or loudness profiles into suggestions like “extend this cut by six frames” or “advance this music cue by 180 milliseconds”? If so, what is gained? and what is lost in such a translation? To even pose this question is to recognize that every signal, every metric, every model is a simplification of the lived experience of cinema. The DST cannot and should not claim to capture “artistic quality.” It can only aspire to optimize specific outcomes: audience engagement, narrative clarity, or memory recall. These are modest, operationalizable goals. Yet they open a space where cinema’s subjective richness intersects with reproducible, falsifiable science.

The first compromise is in the choice of stimuli. To study the cinema is to work with cultural artifacts that are not only technical constructions but also works of meaning. A tool trained on action trailers may fail entirely on art-house documentaries. If we restrict ourselves to short clips in a lab, we gain control and experimental rigor but risk producing results that say little about the lived experience of watching a film. If we show full-length features in naturalistic settings, we capture ecological validity but surrender control: confounds multiply, fatigue sets in, calibration drifts. The DST must therefore live in the in-between. A pragmatic compromise is to use semi-naturalistic sequences, scenes of 6–10 minutes that contain cuts, music cues, dialogue, and emotional beats while validating on longer blocks in more realistic cinema-like environments. This balance acknowledges the impossibility of “pure” control or “pure” realism.

The second compromise is in the metrics. EEG, with its millisecond precision, offers temporal sensitivity but suffers from spatial ambiguity and artifacts. fMRI provides spatial richness but is too slow and expensive to inform editing decisions. Eye-tracking reveals gaze dispersion or clustering but is easily disrupted by calibration drift and cultural bias in models. EDA and HRV capture arousal but blur positive and negative affect, excitement and stress. Every measure is ambiguous. And yet, when EEG synchrony drops, gaze scatter increases, and EDA peaks simultaneously, patterns emerge that are harder to dismiss. The DST does not rely on any single measure but rather on the convergence of signals. This is another philosophical lesson: objectivity in cinema studies cannot mean purity; it can only mean triangulation across partial, noisy, and context-dependent signals.

The third compromise is in causality. Observational data alone can never prove that a neural peak is caused by an editing choice. Peaks may arise from the performance, the dialogue, or simply the brightness of the image. To address this, the DST must embed interventional logic: generate two versions of a scene, one optimized with neurocinematic insights and one left to craft alone, then test them with new viewers. If the optimized edit produces higher recall or more stable attention, we may cautiously infer causal contribution. But even here, humility is necessary. A gain in one metric may correspond to a loss in another: improved clarity may reduce ambiguity, but at the cost of richness; tighter synchrony may increase suspense but diminish openness to interpretation. The DST must never present itself as an arbiter of better art, only as a probe into defined outcomes.

The technical pipeline hides further compromises. Signals must be synchronized, artifacts rejected, baselines defined. Each preprocessing choice however, carries with it the danger of bias. And because every pipeline involves choices, multiverse analyses or testing multiple reasonable paths and reporting whether conclusions hold across them become essential. Here, the philosophical stance is one of radical transparency: better to expose uncertainty than to bury it under false precision.

Even if models prove predictive, the interface to editors is another critical site of compromise. Editors are not neuroscientists. They will not accept black-box outputs that simply say, “Move cut 23 by six frames.” They require explanations, rationales, and confidence intervals. The DST must therefore return interpretable action cards: suggestions framed in terms of both expected effect and underlying reason. “Cut 23 misaligned with neural synchrony; advancing by six frames predicted to increase engagement by 0.12 with 95% confidence.” Such cards resemble color-correction scopes or waveform monitors: diagnostic, not prescriptive. In doing so, the DST acknowledges that optimization is always conditional on creative goals. It resists the reduction of cinema to a single metric of synchrony.

Economics provides its own attack. EEG and eye-tracking sessions cost far more than online surveys or retention analytics. Studios may not justify the marginal gains. Indie filmmakers lack budgets; majors already run test screenings. Here again, compromise points toward targeted use cases. For big-budget trailers, even a 3–5% uplift in retention can justify costs. For educational films or explainers, gains in clarity or recall may be valuable. For art cinema, the DST might function not as an optimizer but as a diagnostic scope, giving editors another lens on their work without prescribing conformity. And as portable EEG, webcam eye-tracking, and remote EDA become cheaper, the cost-benefit balance may shift.

The philosophical takeaway is that the DST is not an attempt to mechanize cinema but to create a new space of dialogue between signals and stories, between data and craft. It acknowledges the limits of objectivity: that neural and audiovisual features are at best proxies for engagement, clarity, or memory. It acknowledges the irreducibility of subjectivity: that audience experience is shaped by culture, mood, and memory, beyond the reach of sensors. And it acknowledges the inevitability of compromise: between experimental control and ecological realism, between causality and correlation, between interpretability and complexity

What survives after all these attacks is not the fantasy of a perfect editing algorithm but the more modest, more defensible claim: that neurocinematic and audiovisual features can, under defined conditions, provide editors with actionable suggestions that measurably affect certain audience outcomes. If the DST can consistently outperform baseline heuristics (motion, loudness, shot length) and traditional audience testing in predicting engagement or recall, then it constitutes an addition to knowledge. If not, it will still have demonstrated, through rigorous failure, the limits of applying neuroscience to cinema. Either outcome advances the field.

The broader philosophical lesson is that cinema itself has always been a negotiation between art and technique. Montage theories from Eisenstein onward have oscillated between claims of scientific universality and acknowledgments of subjective interpretation. Contemporary filmmakers already use waveforms, vectorscopes, and loudness meters—tools that quantify aspects of the audiovisual signal—without believing that these measures determine the meaning of their work. The DST stands in this lineage: a new kind of scope, one that visualizes not the waveform of a sound but the waveform of collective attention. Like all tools, it must be wielded with discretion, humility, and awareness of its blind spots.

My proposal, then, is not to defend a preconceived solution but to invite attack: to let the DST be stress-tested by supervisors, peers, and editors themselves, so that what remains is not fantasy but a durable contribution to knowledge.