If two polygraph questions ask exactly the same thing, why do they not produce exactly the same physiological response?
It is a reasonable question. Many people, including experienced investigators, assume that if two polygraph relevant questions refer to the same alleged behaviour, during the same time period, with nearly identical wording, the body should respond in the same way. If the issue is the same, surely the score should be the same too.
Recent research by Donald Krapohl, Donald Grubin and Ian Dersley shows why that assumption is too simple. Their findings do not weaken professional polygraph science. Properly understood, they show why forensic polygraph examination must be interpreted as a structured psychophysiological assessment rather than as a mechanical word-response machine.
Professional summary: A repeated relevant question is semantically similar, but psychophysiologically new. Expectations, uncertainty, memory activation, physiological recovery and the autonomic nervous system all evolve between the first and second presentation.
The Research Question: Why Do Similar Relevant Questions Diverge?
Krapohl, Grubin and Dersley examined a feature of the British One-Issue Screening Test, or BOST. This format uses two relevant questions about the same core issue. The questions address the same behaviour, within the same time period, and are ordinarily worded in a very similar way. One might therefore expect their scores to be strongly coupled.
The researchers found a more interesting pattern. The two relevant questions were related, but their scores showed only a moderate correlation. Changing the wording of the relevant questions produced little change. By contrast, changing the examiner’s pre-test explanation significantly increased the correlation between the two relevant question scores.
The implication is subtle but important. The difference was not simply in the words themselves. It was in how the examinee understood, anticipated and psychologically organised those words before the test charts were collected. The pre-test interview helped both relevant questions become part of one stable mental representation of the issue.
Polygraph Measures Human Physiology, Not Lies
A polygraph instrument does not detect lies directly. It records physiological activity while a structured sequence of questions is presented. Modern instruments typically record respiration, cardiovascular activity, electrodermal activity and, where used, peripheral blood volume or movement channels.
These channels reflect activity of the autonomic nervous system, the system that regulates breathing patterns, blood pressure, heart rate, sweat gland activity and other involuntary processes. A professional examiner then interprets response patterns using validated formats and an empirical scoring system. The question is not “did the machine find a lie?” The better question is how the person’s physiology changed in relation to carefully formulated question types across repeated charts.
That distinction matters because physiology is not a photocopier. It is a dynamic biological system. Even in controlled laboratory conditions, repeated stimuli rarely produce identical autonomic responses. Perfect correlation between two relevant questions would therefore be biologically surprising.
The Orienting Response: Significance, Not Just Words
The autonomic nervous system is strongly influenced by the orienting response: an automatic shift of attention and physiology towards something novel, uncertain or significant. The response is not triggered only by vocabulary. It is shaped by novelty, expectancy, threat value, personal meaning and the person’s current state of attention.
The first presentation of a relevant question is psychologically different from the second because the first presentation resolves uncertainty. Before it is heard, the examinee may not know exactly how the allegation will be framed, whether unexpected details will appear, or how strongly they will react. Once it has been heard, the question is no longer new in the same way. The brain has updated its model of what is happening.
This helps explain why two relevant questions can ask about the same behaviour yet produce different polygraph scoring values. The semantic meaning may remain stable, while the orienting conditions change.
Habituation and Priming
Repeated stimuli often produce reduced responses, a process known as habituation. People who live beside railway tracks may initially notice every train and later barely register them. A smoke alarm that sounds repeatedly during a drill may provoke a strong first reaction and a weaker later one. Repetition teaches the nervous system that the event is known.
Polygraph testing can involve a similar process. The first relevant question may produce a stronger orienting response because it is the first encounter with that issue in the chart sequence. The second may produce less novelty because the examinee has already heard the issue and survived the immediate moment.
But repetition can also prime a larger response. The first relevant question may activate memory, anxiety, embarrassment, mental imagery or anticipation. By the time the second relevant question arrives, the examinee may be more emotionally prepared, but also more personally engaged. The second response can therefore be smaller, larger or simply different.
Memory Retrieval Is Reconstruction
Modern cognitive neuroscience does not treat memory as a video replay. Remembering is reconstructive. Each act of retrieval can emphasise different details, emotions and associations, and each act of retrieval can slightly alter the neural representation that is later retrieved again.
The hippocampus helps bind together the elements of an autobiographical episode: place, time, people, actions and contextual detail. The medial prefrontal cortex helps relate those details to the self, to social meaning and to longer-term personal narratives. The anterior cingulate cortex is involved in conflict monitoring, error likelihood and the need for control, while the anterior insula contributes to the felt bodily significance of an event, including interoceptive sensations such as tension, arousal or unease.
In plain terms, a relevant question can activate a memory network, a self-relevance network, a conflict-monitoring network and a bodily-feeling network at the same time. The first presentation may emphasise factual retrieval. The second may emphasise anticipated consequence, shame, fear of misunderstanding, or the felt bodily trace of the first response. Repeated relevant questions therefore cannot be assumed to activate identical neural networks.
Consider a person being examined about a concealed affair, a theft from an employer, or an unauthorised disclosure of information. The first relevant question may bring up one detail: a location, a date, a conversation, the person affected. The second may bring up another: fear of consequences, shame, anger, relief that a particular detail was not mentioned, or anxiety that it might appear next.
For a truthful examinee, the same process can occur through fear of false accusation, concern about ambiguity, or uncertainty about how a normal physiological reaction will be interpreted. In both truthful and deceptive examinees, the question is filtered through autobiographical memory and anticipated consequences. The repeated relevant question is therefore not processed by the same brain in the same state.
Salience, Control and Physiological Recovery
Salience is the personal significance of a stimulus at a particular moment. It is dynamic. Fear, embarrassment, confidence, relief, cognitive effort, self-monitoring and anticipation can all shift during an examination. A relevant question that feels routine on one chart may feel highly diagnostic on another.
Examinees also learn throughout testing. They may anticipate when important questions are coming, attempt to regulate their breathing, monitor their body, or mentally prepare for the next presentation. Such executive control can alter respiratory, cardiovascular and electrodermal output even when the question wording remains constant.
Physiological recovery adds another layer. Electrodermal activity may recover more slowly than respiration. Cardiovascular responses have their own timing. Breathing can be affected by the previous question, by posture, or by attempts at control. Each relevant question therefore begins from a slightly different biological baseline.
A Predictive-Processing Account of Score Variability
The original researchers discussed habituation and priming as possible explanations for their findings. A modern way to extend those ideas is through predictive-processing neuroscience. This is a theoretical explanatory model, not something directly measured in the study.
Predictive processing proposes that the brain does not passively receive the world. It constantly predicts what is about to happen and compares incoming information with those expectations. Physiological responses partly reflect the relationship between expected significance, prediction error, precision and uncertainty.
Latent-variable formulation:
P(Rt) = f(S, δt, πt, εt)
- S: underlying issue salience
- A relatively stable representation of the behavioural significance of the investigated issue. It is the enduring personal meaning of what is being examined.
- δt: prediction error
- The discrepancy at time t between what the examinee expected and what was actually experienced during presentation of the relevant question.
- πt: precision weighting
- The confidence, importance or threat value assigned to incoming information. Precision weighting determines how strongly prediction errors influence autonomic responses.
- εt: physiological noise
- Natural biological variability arising from respiration, electrodermal recovery, cardiovascular variability, attention, posture and measurement variability.
This is not presented as a formal scoring equation. It is a conceptual model for thinking about latent variables that may jointly produce an observed physiological response. Bayesian, in this context, does not mean that the examinee is consciously calculating probabilities. It means the brain is continuously combining prior expectations, incoming evidence, uncertainty and personal significance to update its internal model of the examination.
Diagram 1: Predictive Updating Across a Relevant Question
The same words can be processed differently because the prior that receives them has already changed.
Diagram 2: From Latent Variables to Observed ESS Score
The observed empirical score is a summary of channel-level physiology, not a direct reading of truth or deception.
Before the First Relevant Question
Before the first relevant question, the examinee may be uncertain about the wording, the scope of the allegation, whether the examiner possesses undisclosed evidence, how obvious their reaction will be, what happens next and what the consequences may be. That uncertainty can produce a substantial orienting response.
For a deceptive examinee, this may involve concealed memory activation and anticipated discovery. For a truthful examinee, it may involve fear of false accusation and concern about being misunderstood. Predictive processing occurs in both groups. The difference lies in what is being predicted.
After the First Relevant Question
Once the first relevant question has been heard, the brain updates. The examinee now knows the wording, the emotional impact, the expected answer and the fact that the examiner proceeds normally. The second relevant question arrives under different psychological conditions even when it asks essentially the same thing.
This accounts naturally for habituation: prediction error may be smaller because the event is less surprising. It also accounts for priming: the first question may activate memories or fears that make the second question more significant, not less.
Precision Weighting
Predictive-processing theories also use the idea of precision weighting. In simple terms, the brain gives greater weight to information considered important, reliable or threatening. During a forensic polygraph examination, precision may increase because the allegation has been clarified, the consequences are anticipated, the examiner’s explanation has reduced ambiguity, or autobiographical memory has been activated.
Two identical questions can therefore receive different precision weighting. One presentation may feel like another item in a sequence. Another may suddenly feel diagnostic, exposing or consequential.
Hierarchical Prediction During Polygraph Examination
Predictive processing operates across multiple hierarchical levels. A relevant question is not interpreted only at the level of its spoken words. It is interpreted through beliefs about consequences, beliefs about what the examiner knows, beliefs about the behaviour under examination and expectations about the wording of the next question.
Hierarchical Beliefs in a Repeated Relevant Question
Each level continuously updates the others. The autonomic nervous system responds to an evolving hierarchy of beliefs, not simply to spoken words.
A change at a higher level can alter the meaning of a lower-level question. If the examinee begins to believe that the examiner has more information than expected, the same relevant question may acquire greater precision and significance. Conversely, if the pre-test interview clarifies the scope of the allegation and removes feared but irrelevant possibilities, the same words may become less ambiguous and less physiologically disruptive.
Active Inference and the Autonomic Nervous System
Active inference adds another important idea: the nervous system does not merely react to events after they occur. It prepares the body for anticipated events. This is familiar outside polygraph testing. A person may experience a rising heart rate before an important meeting, dry palms before a difficult conversation, or altered breathing before delivering bad news.
In autonomic terms, the body can show anticipatory cardiovascular responses, anticipatory electrodermal activity, respiratory preparation and predictive autonomic regulation. These processes are not necessarily conscious and they are not specific to deception. They reflect the brain’s attempt to prepare the organism for events that are expected to matter.
During a polygraph examination, the first relevant question may teach the examinee when and how the issue appears in the sequence. The second relevant question is therefore encountered by a body that may already be preparing for it. Physiological activity partly reflects prediction rather than passive reaction.
Free Energy and Repeated Relevant Questions
In free-energy accounts of brain function, the brain is often described as attempting to minimise surprise, or more precisely to minimise the discrepancy between its model of the world and incoming evidence. The first relevant question creates uncertainty because it tests the examinee’s predictions about wording, scope, implication and consequence. After that first encounter, the internal model updates.
In many cases, surprise should decrease. The examinee now knows the approximate form of the issue and the fact that the examination continues. This offers a principled account of habituation.
But surprise can also increase. If the first relevant question activates concealed autobiographical memories, intensifies fear of discovery, raises concern about what the examiner may ask next, or makes a previously abstract allegation feel personally vivid, the second relevant question may carry greater prediction error and greater precision. The model therefore naturally predicts cases where RQ2 > RQ1 without implying that the larger second response necessarily proves deception.
The Role of the Pre-Test Interview
This brings us back to the BOST findings. When the examiner’s pre-test explanation made clear that both relevant questions represented the same issue, the correlation between their scores increased. A predictive-processing account would suggest that this reduced ambiguity, reduced uncertainty and helped create one stable mental representation linking R1 and R2.
Higher correlation does not mean identical scores. It means the two scores moved together more consistently. Shared issue salience created stronger coupling, while moment-to-moment prediction error and biological variability still prevented perfect identity.
The Pre-Test Interview as Entropy Reduction
Before testing, the examinee possesses many possible interpretations of the investigation. They may be uncertain about the behavioural scope, the relevant time period, the meaning of key terms, the expected answers, and whether the examiner is asking about one narrow allegation or a wider set of feared possibilities.
The structured pre-test interview progressively removes uncertainty. It defines the behaviour under examination, clarifies the time window, resolves semantic ambiguity, reviews the exact meaning of the questions and establishes the expected yes-or-no answer format. In Bayesian terms, it narrows the examinee’s probability distribution over possible meanings.
The pre-test interview can therefore be understood as an entropy reduction process. It does not remove all anxiety or all physiological variability, but it can reduce the number of competing interpretations that the examinee is trying to evaluate. This may explain why the modified BOST instructions increased correlation between repeated relevant questions: the two questions became more tightly coupled to one common internal representation.
Correlation versus Agreement
Many readers understandably confuse correlation with agreement. Agreement means that two values are the same, or nearly the same. Correlation means that two variables tend to move together.
For example, one relevant question might produce a score of -6 while a second produces -3. Another case might produce +4 and +7. These values differ numerically, so they do not show exact agreement. But across a dataset, paired values that move in the same direction can still show strong positive correlation.
This distinction is important for interpreting empirical scoring systems. The scientific question is not whether R1 and R2 are numerical duplicates. It is whether they are sufficiently related to support interpretation of the underlying psychophysiological pattern.
A Hierarchical Bayesian Model
The latent-variable formulation can be read hierarchically. The observed physiological response is the surface expression of relatively stable issue salience, moment-specific prediction error, precision weighting and biological noise.
P(Rt) = f(S, δt, πt, εt)
Underlying issue salience is the relatively stable personal significance of the matter under investigation. Prediction error captures moment-to-moment differences in expectation, interpretation and emotional preparation. Precision weighting determines how strongly those prediction errors influence autonomic output. Physiological noise includes respiration, electrodermal recovery, cardiovascular variability, attention, posture and measurement variability.
Shared salience creates correlation. Prediction error, precision weighting and biological variability prevent perfect agreement. This is why score variability is not a methodological embarrassment. It is what one should expect when measuring living psychophysiology.
Scientific Limitations
The predictive-processing account is plausible, but it remains a hypothesis. The study did not directly measure subjective uncertainty, prediction error or precision weighting. Future research could examine question order, subjective expectation, ground-truth cases, channel-specific habituation, memory retrieval and the degree to which pre-test explanations alter psychological coupling between repeated relevant questions.
That limitation should be stated clearly. The value of the model is that it integrates orienting responses, salience, habituation, priming, memory reconstruction, autonomic physiology and biological variability into one scientifically coherent framework. It does not prove that this framework is the final explanation.
Bayesian Evidence Accumulation and Empirical Scoring
The examinee is not the only person updating a model. Across the examination, the examiner is also accumulating evidence about the stability, consistency and interpretability of the physiological pattern. Repeated observations matter because they allow a professional to distinguish a robust pattern from an isolated response, artefact or transient physiological fluctuation.
In this limited sense, empirical scoring resembles Bayesian evidence accumulation. Each chart contributes information. Each physiological channel contributes information. Repetition increases confidence when observations converge, and reduces confidence when they are unstable, contaminated by artefact or inconsistent with the expected pattern.
This is a conceptual analogy, not a claim that empirical scoring systems are formally Bayesian. The point is more modest: professional interpretation is cumulative. It evaluates how much confidence should be placed in the total psychophysiological pattern rather than treating any single relevant question response as decisive.
Why Professional Examiners Use Multiple Charts
No competent forensic polygraph conclusion should rest on one isolated reaction to one isolated question. Professional practice uses multiple charts, multiple physiological channels, structured question formats, empirical scoring and aggregation across observations. Averaging repeated observations improves reliability because random biological variation is less influential than the overall pattern.
This is also why the testing environment, question formulation and pre-test procedure matter. They reduce unnecessary sources of variability and make the recorded data more interpretable.
Practical Implications
The finding that two relevant questions can produce different scores should not surprise us. Variation is expected. Professional forensic polygraph scoring systems are designed for that reality. Interpretation depends on the full pattern of psychophysiological evidence rather than on the assumption that the same words must produce the same response every time.
Indeed, the pre-test finding strengthens confidence in professional examinations. It shows that careful explanation, reduced ambiguity and scientifically formulated relevant questions can make the examinee’s psychological processing more consistent. That is professional methodology in action.
Conclusion
A repeated relevant question is not simply another presentation of identical words. It is a new interaction between an evolving brain, an evolving autonomic nervous system and an evolving internal model of the examination. The question may be semantically identical, yet the psychophysiological context has changed.
Appreciating that distinction moves forensic polygraph interpretation beyond simplistic stimulus-response thinking towards a contemporary neuroscientific understanding of human behaviour. It also explains why professional forensic polygraph interpretation evaluates the entire psychophysiological pattern, rather than expecting identical questions to produce identical numerical scores.
Professional Practice at The Centre for Forensic Neuroscience
Examinations conducted by Dr Keith Ashcroft employ validated examination formats, structured pre-test interviews, scientifically formulated relevant questions, multiple physiological recording channels, recognised empirical scoring methods and careful consideration of the examinee’s psychological state.
Conclusions are based on the overall pattern of psychophysiological evidence rather than isolated physiological reactions. Modern forensic polygraph examination is not a claim to infallibility and it is not a machine-based lie detector. It is a structured psychophysiological assessment grounded in decades of research into attention, salience, memory and autonomic nervous system functioning.
Further Reading
- Friston, K. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience, 11, 127-138.
- Clark, A. (2016). Surfing Uncertainty: Prediction, Action, and the Embodied Mind. Oxford University Press.
- Hohwy, J. (2013). The Predictive Mind. Oxford University Press.
- Seth, A. K. (2013). Interoceptive inference, emotion, and the embodied self. Trends in Cognitive Sciences, 17(11), 565-573.
- Krapohl, D., Grubin, D. and Dersley, I. (2026). Recent study of repeated relevant-question scoring in the British One-Issue Screening Test.
- Nelson, R. (2015). Scientific basis for polygraph testing. Polygraph.
This article is provided for general information and educational purposes. It does not constitute legal advice, clinical advice or a substitute for professional consultation. Polygraph results should always be interpreted cautiously and in context by qualified professionals.
Frequently Asked Questions
What are polygraph relevant questions?
Relevant questions are the issue-focused questions in a structured polygraph examination. They refer to the behaviour or event under investigation and are compared with other question types using recognised scoring methods.
Why can two similar relevant questions produce different polygraph scores?
The semantic wording may be similar, but the examinee's psychological state is not identical from one presentation to the next. Attention, expectation, memory retrieval, salience, physiological recovery and biological variability can all change between questions.
Does score variability mean polygraph scoring is unreliable?
No. Some variability is expected in human physiology. Professional interpretation relies on multiple charts, multiple physiological channels, empirical scoring and the overall pattern of responses rather than a single isolated reaction.
What role does the pre-test interview play?
A structured pre-test interview clarifies the issue, reduces ambiguity, reviews the relevant questions and helps the examinee understand what each question represents. This can improve the psychological consistency with which repeated relevant questions are processed.
Is predictive processing proven as the explanation for repeated relevant question variability?
No. Predictive processing is a plausible theoretical framework that helps integrate orienting responses, habituation, priming, memory reconstruction and autonomic physiology. It remains a testable hypothesis rather than a direct finding of the study discussed.