Unveiling Deceptive Intentions: Insights from Eye Movements and Pupil Size

This study examined eye movements and pupil responses during different types of deception using a card-based Concealed Information Test paradigm. The research distinguished between faking (actively presenting false information) and concealing (hiding truth) behaviors to identify distinct physiological signatures of deceptive intentions.

87 participants performed a card selection task under three conditions (revealing truth, faking, concealing) while eye movements were tracked at 1000Hz using EyeLink technology in a controlled laboratory setting.

This research applied eye-tracking technology to deception detection within a Concealed Information Test framework, addressing the gap between different types of deceptive behavior. Using a card game paradigm, 87 participants were instructed to fake, conceal, or tell the truth while their eye movements and pupil responses were monitored at high resolution. The study found distinct fixation patterns between concealing versus other conditions, with pupil size additionally revealing deceptive objects. The number and duration of fixations emerged as the most reliable indicators of deceptive intentions, suggesting that different forms of deception produce distinct cognitive loads and physiological signatures.

This research advances non-invasive deception detection by demonstrating that eye-tracking can differentiate between types of deceptive behavior, potentially offering a more nuanced and less intrusive alternative to traditional polygraph techniques for detecting concealed information.

Background & Context

The quest to detect deception through observable physiological cues has intrigued researchers for generations, yet identifying reliable indicators of lying remains one of psychology's most elusive challenges. While recognition of deceptive intentions from the eyes has been of appealing interest in the last decades, it is still unresolved. Traditional polygraph techniques measure cardiovascular, respiratory, and electrodermal responses, but eye-tracking technology offers a novel, potentially less invasive approach to deception detection that capitalizes on involuntary ocular responses.

This research addresses the unresolved question of recognizing deceptive intentions from the eyes by developing a paradigm based on the Concealed Information Test enabling the study of various kinds of deception—specifically faking and concealing. The study fills a critical gap by distinguishing between different forms of deception rather than treating all lies as equivalent. Previous research often conflated various deceptive behaviors, but emerging evidence suggests that concealing information (hiding the truth) and faking information (actively presenting false information) may produce distinct cognitive loads and therefore different physiological signatures.

This work represents an important advancement in applying eye-tracking methodology to deception detection within the framework of the Concealed Information Test (CIT), a scientifically validated alternative to traditional polygraph techniques that examines whether individuals possess crime-relevant knowledge. By examining both fixation patterns and pupillary responses simultaneously, the researchers sought to create a more comprehensive understanding of how different types of deceptive intentions manifest in ocular behavior.

Research Design & Methodology

Based on a card game, researchers compared fixation as well as pupil behavior while participants were instructed to fake, conceal, or tell the truth, and realized two different layouts of stimulus presentation. The dataset contains 87 participants, including 56 females with a mean age of 22.1 years (SD=2.1 years). Participants selected a card numbered 1 through 6, then viewed all possible card values presented on screen while their eye movements were monitored.

Eye movements were measured using the eye tracker EyeLink 1000 from SR Research Ltd, with a 1000Hz sample frequency, with participants seated at a distance of 57cm from a screen in a room with low ambient light, with their heads on a chin rest to prevent noise movements. The experimental design incorporated three distinct instruction conditions that created different deceptive contexts:

• Revealing (truth-telling): Participants had to disclose the target to a "partner" interlocutor, wanting the partner to realize which number they selected
• Faking: Participants had to mislead an "opponent" interlocutor into guessing another card than the target, with the choice of fake number up to the participant
• Concealing: Participants had to prevent a "referee" interlocutor from guessing the target

Experiments contained the selection of a target, followed by the presentation of a baseline card and the number cards one after another for 5s each, with number cards separated by baseline cards for 2s in the Eyelink dataset. The two stimulus layouts tested whether presentation format (one card at a time versus multiple cards simultaneously) influenced deceptive gaze patterns, addressing inconsistencies in prior literature.

Results & Key Findings

Fixations differed between concealing versus faking and telling the truth; pupil size additionally unveiled the object of deception, with the best candidates to reveal deceptive intentions being the number and duration of fixations, while saccade amplitudes and pupil size showed great potential. The study demonstrated that different types of deception produce distinct ocular signatures that can be reliably measured and differentiated.

When cards were presented sequentially one at a time, concealment was associated with fewer and longer fixations, whereas when six cards were presented simultaneously, concealment produced more and shorter fixations. This striking reversal reveals that stimulus layout fundamentally alters the cognitive strategies participants employ when deceiving, with critical implications for interpreting eye movement data in operational settings.

Key findings from the eye-tracking measures included:

• Fixation patterns: Target presentations produced more and shorter fixations, fewer blinks, and more gaze transitions between cards in the concealing and faking-target conditions compared to revealing and faking-faked-target conditions
• Pupil size differentiation: Pupillary responses successfully identified which specific stimulus was the concealed target, going beyond simple detection of deceptive intent to reveal the actual object of deception
• Saccade measurements: Among the 60 eye movement features investigated, saccade number, duration and amplitude along with maximum pupil size were identified as the most important predictors of deception

Subsequent machine learning analysis of this dataset achieved detection accuracies that varied by task complexity. Detection of card concealment achieved accuracies ranging from 64% to 74% in binary classification, while the three-classification task proved more complex with 49% accuracy, demonstrating both the promise and current limitations of automated deception detection systems.

Discussion & Significance

This research makes a foundational contribution by demonstrating that various kinds of deceptive behavior must be carefully distinguished, contributing to how to use gaze measures as indicators of deceptive intentions. The finding that concealing and faking produce different eye movement patterns challenges the widespread assumption in deception research that all lies are cognitively equivalent. This distinction has profound theoretical implications for understanding the neural and cognitive mechanisms underlying deception.

The pattern suggests that simultaneous object presentation elicits qualitatively different deceptive eye movements than sequential presentation, a factor that is critical for the interpretation of eye movements in CIT paradigms and may contribute to inconsistent findings across previous studies. This discovery helps resolve contradictions in the literature where some studies found increased fixations during deception while others found decreased fixations—the presentation format likely explains these divergent results.

Unlike intrinsic deception-induced cognitive load arising from truth inhibition, strategic modulation originates from high-order cognition and reflects deliberate eye movement strategies adapted to the stimulus configuration, indicating that fixation behavior in deception contexts reflects the combined influence of deception-induced cognitive load and strategic gaze modulation. This nuanced understanding moves beyond simplistic cognitive load theories to recognize that deceivers actively manage their gaze behavior, sometimes successfully concealing their intentions through strategic eye movements.

Limitations & Considerations

The study employed a controlled laboratory paradigm with relatively low-stakes deception involving card games, which differs substantially from real-world interrogation contexts where consequences are severe and emotional arousal is heightened. The sample consisted primarily of young adults (mean age 22.1 years) with normal or corrected-to-normal vision and the absence of epileptic seizures, limiting generalizability across age groups and populations with visual impairments or neurological conditions.

Previous research showed that Revealing and Faking tasks in a CIT induce similar eye movements, and while AI enhances differentiation, achieving a clear distinction between these behaviors remains challenging. The chin rest requirement, while necessary for measurement precision, creates an artificial constraint absent in real interrogations. Additionally, participants knew they were being monitored and that deception was expected, potentially reducing ecological validity compared to situations where individuals spontaneously choose to deceive.

Practical Applications

For polygraph examiners and forensic investigators, this research provides actionable insights about the importance of standardized stimulus presentation formats when incorporating eye-tracking into Concealed Information Tests. The finding that fixation patterns reverse based on whether stimuli are presented sequentially versus simultaneously means that testing protocols must be carefully designed and consistently applied. Examiners should be trained to recognize that concealing known information produces different ocular signatures than actively fabricating false information, requiring different interpretive frameworks.

The identification of specific eye parameters—fixation number and duration, saccade characteristics, and pupil size—as reliable indicators offers a roadmap for developing next-generation deception detection systems. However, the moderate accuracy rates achieved (64-74% for binary classification) indicate that eye-tracking should augment rather than replace traditional polygraph measures. For consumers of lie detection services, these findings underscore that not all deception detection methods are equivalent and that the specific testing paradigm matters significantly for result interpretation.