Assessing partial errors via analog gaming keyboards in response conflict tasks: A proof-of-concept study with the concealed information test

This proof-of-concept study explored using analog gaming keyboards to detect response conflict in the Concealed Information Test through partial button presses. Results showed that probes produced significantly more partial errors than irrelevant items, providing a new accessible method for measuring response conflict without specialized equipment.

Participants (n=35) completed an autobiographical RT-CIT using analog gaming keyboards while concealing their identity, with validation through a modified Sternberg task (n=33).

This study investigated whether commercially available analog gaming keyboards could measure response conflict in the Concealed Information Test by detecting partial button presses. Participants concealed their identity while using sensitive keyboards that could detect incomplete motor responses. The research found that concealed information (probes) produced significantly more partial errors than control items, demonstrating that gaming keyboards can successfully detect response conflict without specialized physiological equipment. The findings were validated using a modified Sternberg recognition task, showing consistent results across paradigms.

This technological innovation could democratize access to sophisticated response conflict measurement in deception detection, making advanced CIT research and applications more accessible without requiring expensive specialized equipment or trained physiological recording technicians.

Background & Context

The response time-based Concealed Information Test (RT-CIT) has emerged as a promising behavioral alternative to traditional polygraph testing, relying on the principle that slower response times to critical information (called 'probes') compared to control items (called 'irrelevants') reveal recognition. Research has increasingly focused on understanding the underlying psychological mechanisms driving this effect, with response conflict identified as a strong contributor to this RT difference.

Previous attempts to measure response conflict in the CIT relied on electromyography (EMG), which requires special equipment and trained examiners. This created a significant barrier to both research advancement and practical application. The 2024 study by Koller, Hofer, and Verschuere addresses this methodological gap by exploring whether commercially available analog gaming keyboards—designed to detect even minimal finger pressure—could serve as an accessible alternative for measuring response conflict through "partial errors" (incomplete button presses of the incorrect response key).

This proof-of-concept research represents an important technological innovation in deception detection methodology, potentially democratizing access to sophisticated response conflict measurement without specialized physiological recording equipment. The study was funded by the Swiss Federal Office of Civil Aviation and the Zurich State Police, Airport Division, highlighting practical law enforcement interest in improving memory detection techniques.

Research Design & Methodology

In a preregistered study, participants completed an autobiographical RT-CIT (n = 35) as well as a cued recognition task (modified Sternberg task; n = 33) for validation purposes. The autobiographical RT-CIT employed a concealed identity scenario where participants were instructed to hide their true identity and pretend to be someone else. Participants were asked to imagine fleeing a country while carrying a fake ID, then being stopped by border control at the airport and tested for their identity.

The experimental design incorporated several key elements:

• Participants used analog gaming keyboards sensitive to partial key presses, measuring incomplete motor responses
• Participants pressed YES when presented with fake ID information (targets) and NO for all other information (irrelevant items and probes)
• Memory for fake identity details was verified through free recall testing, with only error-free participants proceeding to the RT-CIT
• A modified Sternberg task served as validation, testing whether partial errors occurred in a well-established conflict paradigm

Analyses were conducted in R with BayesFactor and brms packages, employing Bayesian mixed effects ANOVAs with JZS priors. The researchers excluded target trials, trials with response times outside 200-1500 ms, response errors, and trials beginning with partially pressed keys, resulting in 1.82% of probe and irrelevant trials being excluded.

Results & Key Findings

The study provided strong evidence that analog gaming keyboards can successfully detect response conflict in concealed information testing. The CIT effect was apparent in the frequency of partial button presses, with a Bayes Factor of 275 providing strong evidence for an effect of item type.

Key statistical findings included:

• Probe trials showed 2.93% partial errors (SD = 3.73%) versus 0.46% for irrelevant items (SD = 0.88%)
• In the Sternberg validation task, intrusion trials versus new trials yielded a Bayes Factor of 102
• Overall, partial errors remained rare at 1.3% of valid trials in the Sternberg task (intrusion: 1.71%, new: 0.16%, match: 1.72%)
• The effect demonstrated consistency across two different cognitive tasks, supporting the generalizability of the measurement approach

While partial errors occurred significantly more frequently during response conflict, they remained relatively uncommon overall—appearing in less than 3% of conflict trials. This low base rate suggests the measure captures subtle motor activation that traditional accuracy measures miss entirely, providing additional diagnostic information beyond response time and explicit errors.

Discussion & Significance

This research represents the first empirical demonstration that commercially available analog keyboards can measure partial motor responses indicative of response conflict in deception detection contexts. The findings have important theoretical implications for understanding the cognitive mechanisms underlying the CIT effect, providing direct behavioral evidence that examinees experience motor conflict when attempting to conceal recognition of familiar information.

The study advances a growing body of research on response conflict as a core mechanism in memory detection. Unlike EMG equipment, analog gaming keyboards require no electrode placement, no specialized technical expertise, and cost a fraction of traditional psychophysiological recording systems. This accessibility could enable wider research participation, larger-scale validation studies, and potentially more practical deployment scenarios where full EMG setup is impractical.

The convergence of findings across both the RT-CIT and the modified Sternberg task strengthens confidence that the keyboard methodology captures genuine cognitive conflict rather than task-specific artifacts. The Bayesian analytical approach, with Bayes Factors exceeding 100 in both paradigms, provides robust quantification of evidence strength—a methodological improvement over traditional frequentist approaches that only permit rejection or non-rejection of null hypotheses.

Limitations & Considerations

The researchers acknowledge several important constraints on interpretation. Most significantly, partial errors remained rare even during conflict trials (2.93% for CIT probes, 1.71% for Sternberg intrusions), suggesting the measure may lack sensitivity for individual-level classification. While the group-level effects were robust, the low frequency of partial errors limits their practical utility as a standalone diagnostic indicator in applied settings.

The study employed an autobiographical concealed identity paradigm with motivated participants in a laboratory setting. Generalization to real-world forensic contexts—where stakes, stress levels, and examinees' cognitive strategies may differ substantially—remains uncertain. Additionally, while analog keyboards proved viable for measuring partial errors, the authors note they are likely less sensitive than EMG in detecting subtle muscle activation patterns.

The research focused exclusively on response conflict measurement without examining whether partial error data improve classification accuracy beyond standard RT-CIT metrics. Future validation studies would need to demonstrate added diagnostic value before recommending incorporation into operational protocols.

Practical Applications

For polygraph professionals and researchers, this study opens new possibilities for accessible response conflict assessment in CIT applications. The technology requires minimal investment and no specialized physiological monitoring skills, potentially enabling screening applications in airports, border control, or employment contexts where EMG deployment would be cost-prohibitive or logistically impractical. The keyboard approach could also facilitate large-scale online research studies investigating deception mechanisms.

However, consumers and practitioners should recognize that this proof-of-concept study demonstrates measurement feasibility rather than diagnostic readiness. The low frequency of partial errors and absence of validation against ground-truth outcomes in applied contexts means the technique is not yet ready for operational decision-making. The research contributes to theoretical understanding of CIT mechanisms and provides researchers with a new tool for investigating response conflict, but substantial additional validation work is needed before partial error measurement could augment operational concealed information testing protocols.