An N400 response is a characteristic negative-going pattern in an event-related brain potential, measured using EEG, that is sensitive to the relationship between a meaningful stimulus and its context. The name refers to a negativity often prominent around 400 milliseconds after a target. Researchers commonly find larger N400 negativity for words that are less semantically supported by their preceding sentence, although the effect is more complex than a simple detector of mistakes.
Did you know that a word can be perfectly grammatical yet feel very unexpected in the sentence that contains it? The invented sentence ‘She spread butter on the warm cloud’ gives readers a different interpretive challenge from ‘She spread butter on the warm bread’. N400 research helps scientists measure how context influences meaning processing. For students, the discovery points to a broader truth: vocabulary is not a collection of isolated entries but a system that helps interpret each word in relation to the ideas around it.
This guide explains the original 1980 findings of Marta Kutas and Steven Hillyard, what EEG and ERPs measure, why N400 amplitude relates to contextual expectancy and association, and what researchers still debate about meaning access and integration. It includes original applications for Primary and Secondary English, Mathematics, Science, History, General Paper and translation, along with safeguards against mistaking neural processing for factual truth or learner ability.
Research boundary: All classroom contrasts below are illustrative. No EEG has been recorded for this article, and no voltage, participant score or individual diagnosis is invented.
Contents: what a word means depends on where it appears
Open 44 explanations, research safeguards and teaching cases
- 1. The N400 is a response connected with meaning
- 2. Vocabulary gives expectations a foundation
- 3. Kutas and Hillyard’s 1980 discovery
- 4. What ERP means
- 5. Why the name includes N and 400
- 6. The comparison is often a difference in amplitude
- 7. A fictional sentence-ending contrast
- 8. Oddness does not automatically mean ungrammaticality
- 9. A correct sentence can contain a surprising word
- 10. Cloze probability provides an independent estimate
- 11. The 1984 expectancy finding matters
- 12. Relatedness and probability are distinct
- 13. N400 is not restricted to written English
- 14. A picture can be semantically unexpected
- 15. Spoken language unfolds differently from print
- 16. The scalp distribution is not a simple brain map
- 17. Many trials are needed for a useful average
- 18. The baseline and reference matter
- 19. Blinking can distort scalp signals
- 20. A component is not a physical packet of meaning
- 21. Access and integration are research explanations
- 22. N400 is not a literal surprise detector
- 23. N400 and P600 are not interchangeable
- 24. Grammar can influence semantic processing
- 25. Negation changes the proposition
- 26. Word frequency can affect the experiment
- 27. Sentence position can alter interpretation
- 28. Prediction is not the only source of meaning
- 29. The same word can have several senses
- 30. A metaphor can be coherent without literal features
- 31. Context can make unusual sentences meaningful
- 32. A false fact can sound familiar
- 33. Clinical research has a different purpose
- 34. Ethical EEG research needs informed participation
- 35. A high-school explanation can use two sentences
- 36. Primary English: identify the contextual clue
- 37. Secondary English: avoid automatic substitutions
- 38. Science: a surprising fact still deserves testing
- 39. Mathematics: language controls the operation
- 40. History: familiar narrative can obscure evidence
- 41. General Paper: coherence is not enough
- 42. Translation: preserve the intended sense
- 43. A reproducible research notebook
- 44. The final lesson: meaning is constructed in context
Quick test: meaningful context and an unlikely ending
| Original illustrative sentence | Why it matters | What cannot be claimed |
|---|---|---|
| She spread butter on the warm bread. | The completion fits an ordinary food preparation context. | No measured N400 amplitude is available for this invented sentence. |
| She spread butter on the warm cloud. | The same preceding context supports a less ordinary ending. | It is not necessarily ungrammatical or false within every fictional setting. |
| The gardener watered the flowers. | A familiar action predicts relevant objects. | Prediction is not proof that any particular garden contains flowers. |
| The gardener watered the robot. | An unusual object may require added story context. | No experiment here establishes an individual learner’s neural response. |
This table demonstrates why researchers need controlled sentence sets and independent norms. A teacher can discuss meaning expectations without claiming the class produced an electrical response or clinical measure.
1. The N400 is a response connected with meaning
The N400 is a characteristic pattern in an event-related potential, or ERP, recorded from the scalp while people process meaningful stimuli. It commonly appears as a relatively negative-going response whose activity is prominent around 400 milliseconds after the relevant stimulus, though its onset, distribution and timing vary across tasks.
It is not a brain alarm that announces every false statement. Researchers study how the N400 changes between controlled conditions. Words that fit less readily with their preceding context often elicit larger N400 negativity than closely expected or semantically supported alternatives. That difference tells us about processing, not about whether a pupil is clever.
2. Vocabulary gives expectations a foundation
A listener or reader has learned patterns connecting words, grammatical constructions and real-world concepts. When the sentence begins ‘The gardener watered the …’, knowledge of gardens and familiar objects makes some completions more plausible than others.
The N400 offers one window into how those expectations and conceptual associations affect processing. A strong vocabulary system allows readers to connect words to situations and distinguish plausible meanings. The effect is not proof that a particular completion is true; expectations and evidence must still be checked.
3. Kutas and Hillyard’s 1980 discovery
Marta Kutas and Steven Hillyard reported an influential experiment in Science in 1980 titled Reading Senseless Sentences: Brain Potentials Reflect Semantic Incongruity. Sentence endings that did not fit the preceding context produced a distinctive negative ERP response.
Their result helped establish ERPs as tools for investigating language comprehension. The original experiment is more specific than a claim that the brain simply detects ‘bad English’. A semantic anomaly, physical visual change and grammatical violation can influence different aspects of the recorded signal.
4. What ERP means
An event-related potential is a signal derived from electrical recordings of the brain, usually by analysing electroencephalography, or EEG, around repeated events. Researchers align the recorded signal to stimulus onset and compare averaged responses across trials.
The ERP is a tiny voltage pattern observed at scalp electrodes; it is not a photograph or direct measurement of an individual word’s stored meaning. Interpreting a component requires knowing the task, event timing, electrode configuration and comparison condition.
5. Why the name includes N and 400
The letter N refers to a negative-going component in the conventional ERP naming scheme, while 400 refers approximately to its prominent latency in milliseconds. Naming conventions are historical, and the actual measured response spans a time interval rather than a single instantaneous event.
A child should not be told that every brain understands a word precisely at 400 milliseconds. The N400 can begin earlier, continue beyond the peak and vary with sensory modality and context. Its value lies in the systematic contrast, not a mythical universal stopwatch.
6. The comparison is often a difference in amplitude
Researchers commonly compare average ERP amplitude elicited by target words in different contexts. Under many paradigms, an unexpected or semantically incongruous word produces a more negative N400 than a highly supported word.
The direction depends on signal plotting conventions, reference choices and the condition contrast. A larger N400 is not a higher grade, a personal level of intelligence or a direct count of how many definitions someone remembers.
7. A fictional sentence-ending contrast
Consider the original pair ‘She spread butter on the warm …’ followed by bread versus cloud. For many readers, bread fits the everyday event more readily; cloud creates an unusual interpretation. Such pairs illustrate why researchers compare expected and unexpected completions.
These invented sentences have not been recorded with EEG or normed for a participant population. They illustrate a hypothesis: under a suitable design, contextual support could affect the relative N400. No voltage difference or experiment outcome is claimed here.
8. Oddness does not automatically mean ungrammaticality
A sentence can be grammatically well formed while semantically unusual. ‘The pillow negotiated with the traffic light’ has a recognisable subject and verb structure but presents an unexpected ordinary-world event.
Conversely, a sentence can express a comprehensible idea while containing grammatical errors. The N400 research tradition first became well known through semantic anomalies, but its sensitivity is broader than this simple classroom contrast. Separate grammar, meaning and factual status.
9. A correct sentence can contain a surprising word
A person might be told about an unfamiliar tool, rare animal or extraordinary but documented event. The final word can be difficult to predict while remaining accurate and meaningful once the full context is known.
Therefore, unexpectedness is not proof of falsity. In Science and History, a surprising fact may be the most valuable fact in the passage. A responsible reader learns the new concept rather than rejecting it because the wording did not match their initial expectation.
10. Cloze probability provides an independent estimate
A common research measure of a word’s predictability is cloze probability. Researchers present an incomplete sentence to participants and ask for the word they expect next. The proportion supplying a particular completion estimates that completion’s predictability for that group.
Kutas and Hillyard’s 1984 study related N400 amplitude to word expectancy as measured by cloze probability. The calculation is a behavioural norm, not a neural readout. A participant can choose an unlikely completion without having abnormal language processing.
11. The 1984 expectancy finding matters
In their Nature report, Kutas and Hillyard showed that N400 amplitude varied with the expectancy of sentence-final words. They also observed effects of semantic relatedness even for unexpected completions.
This complicates the story that the N400 merely marks meaningless words. A completion that is unexpected but related to the likely idea may be processed differently from a wholly unrelated one. Vocabulary networks influence what counts as a supportive context.
12. Relatedness and probability are distinct
An item may be semantically associated with an expected word without being a common sentence completion. For example, butter may be strongly related to bread, yet not be grammatically appropriate at every position in a sentence about bread.
A rigorous experiment separates how frequently participants supplied the target from how strongly it relates to a different likely item. Association, plausibility and actual completion probability are different properties.
13. N400 is not restricted to written English
Kutas and Federmeier’s 2011 review discusses N400 responses to written and spoken words and to meaningful nonlinguistic stimuli, including pictures and meaningful actions. This extends its role beyond simple errors at the end of printed sentences.
For vocabulary education, the broader finding is that meaning connects across representational formats. A student may recognise a concept through speech, print and pictures. That relationship should not be exaggerated into a claim that all modalities produce identical signals.
14. A picture can be semantically unexpected
If a visual scene strongly establishes one activity, a later object inconsistent with that activity may change semantic processing. ERP studies use such comparisons to investigate how concepts integrate with visual context.
The N400 is therefore not an English spelling checker. Meaning can be activated without alphabetic letters. Teachers can use real examples to connect words with objects and actions, while remembering that classroom observation is not laboratory EEG evidence.
15. Spoken language unfolds differently from print
A printed word can be presented at a specific instant, while speech develops through time. The onset of an auditory word may leave several lexical candidates possible before later sound narrows them.
N400 studies using speech require appropriate alignment to acoustic events and attention to recognition timing. An ERP peak relative to word onset does not imply every spoken word became unambiguous at the same moment.
16. The scalp distribution is not a simple brain map
ERP researchers record electrical potentials at multiple scalp electrodes and may describe where a component is most prominent. A scalp voltage pattern does not by itself identify one unique brain generator or precise anatomical location.
Students should resist headlines claiming that an electrode above one region shows a vocabulary word stored directly below it. Neural sources need appropriate modelling and converging evidence. The spatial resolution of scalp EEG has important limits.
17. Many trials are needed for a useful average
A single trial contains neural activity, eye movements, muscle signals and measurement noise. ERP experiments typically repeat controlled conditions and average responses to improve the signal-to-noise ratio.
This makes an impressive waveform a product of many observations and methodological decisions. A teacher cannot interpret one spontaneous facial reaction as an N400. Without suitable recording and controls, the educational activity remains an analogy.
18. The baseline and reference matter
EEG voltage is measured relative to a reference, and ERP studies often use a baseline interval for comparison. The choice of reference, filtering and baseline correction can affect the reported waveform.
A trustworthy article reports relevant methodological details instead of showing a decorative negative peak. Readers need to know which conditions were compared and how the amplitude was measured. Precision in the method protects precision in the claim.
19. Blinking can distort scalp signals
Eye blinks and other movements produce electrical artifacts in EEG that may be much larger than language-related signals. Researchers use appropriate recording procedures, artifact detection and analysis rules to deal with these sources.
A sample should not be declared abnormal because of one noisy trace. Data exclusion and preprocessing decisions need documentation. Vocabulary is not measured by placing an ordinary headset on a pupil and guessing from visual fluctuations.
20. A component is not a physical packet of meaning
The label N400 identifies an experimentally observed response pattern related to semantic processing. It does not mean that a single word’s dictionary definition is transported as one isolated electrical pulse.
Theoretical accounts of the N400 discuss semantic access, activation, integration and prediction. These mechanisms may overlap in some models and differ in others. A scientifically careful guide should distinguish the reproducible measurement from debates over its exact function.
21. Access and integration are research explanations
One account emphasises the accessibility of conceptual information activated by a preceding context; another emphasises difficulty integrating new information into the evolving interpretation. Research has tested versions of both ideas.
Kutas and Federmeier’s review highlights dynamic interplay between bottom-up and top-down information. A student need not choose a single oversimplified theory to benefit from the concept. The core finding is that the context changes measurable meaning processing.
22. N400 is not a literal surprise detector
A reader can be surprised by a change in font, an unexpected colour or a sudden sound without the same semantic effect as an unexpected word. Physical novelty and semantic relatedness are separable variables.
The original Kutas and Hillyard experiments compared linguistic and physical deviations. Consequently, a large N400 should not be described as the universal brain response to any surprise. The nature of the stimulus and comparison matters.
23. N400 and P600 are not interchangeable
P600 is a later positive-going ERP effect often studied in relation to syntactic anomaly or reanalysis, while the N400 is a negative-going effect strongly associated with meaning processing. Both labels describe experimental signals rather than school error codes.
Research also shows that semantically anomalous sentences can elicit P600 effects under some conditions. Thus the simple slogan ‘N400 means meaning and P600 means grammar’ is only a beginner’s rough distinction, not a universal decoding rule.
24. Grammar can influence semantic processing
Sentence grammar helps determine who is doing what to whom. The word order and syntactic roles of a sentence affect which interpretation is built from otherwise familiar words.
For example, ‘The dog chased the child’ differs from ‘The child chased the dog’ despite using the same content words. An N400 study must consider context and structure, while an educational lesson requires the pupil to explain the actual event.
25. Negation changes the proposition
A short word such as not can reverse the meaning of a sentence. ‘The water is safe’ and ‘The water is not safe’ share many prominent words but describe different conditions.
N400 research cannot be turned into an automatic truth-checking service for such statements. Students must read the negation, identify the current source and verify the factual claim. Frequency of a word does not determine its logical importance.
26. Word frequency can affect the experiment
High-frequency words are often encountered more regularly than low-frequency words in particular corpora, and this can influence recognition and processing. If expected items are also more familiar than unexpected ones, a study may have more than one changing variable.
Research designs match or model relevant lexical factors. A descriptive result about a particular sentence should not be overgeneralised to every learner. Corpus frequency is one property, not the entire explanation of comprehension.
27. Sentence position can alter interpretation
Words at the beginning, middle and end of sentences have different surrounding context available. A sentence-final item may be strongly constrained by earlier words, while the first item arrives before much linguistic context exists.
A well-designed ERP study compares positions appropriately or makes its restriction explicit. A pupil should not assume that an effect reported for sentence-final completions applies without change to all possible positions.
28. Prediction is not the only source of meaning
A reader can understand a word after encountering it even if it was not expected beforehand. Understanding involves interaction between the form, stored lexical knowledge, context and sometimes newly supplied explanations.
Educationally, an unfamiliar technical term can be introduced deliberately and then defined. The fact that its first appearance was surprising is not a failure of teaching. The learning question is whether the new concept becomes usable later.
29. The same word can have several senses
A context about money may favour the financial sense of bank, while a river description favours the geographical sense. The word’s spelling alone cannot establish which concept is intended.
A pupil should compare the neighbouring words and provide an appropriate paraphrase. A neural response to an ambiguous target can be informative in research, but a classroom success is a justified interpretation, not a claimed waveform.
30. A metaphor can be coherent without literal features
Sentences such as ‘Her criticism cut deeply’ contain figurative uses of familiar words. An interpretation based only on literal object properties would fail to explain the meaning.
An N400 is not guaranteed whenever a verb is metaphorical. Conventionality, context and familiarity matter. Teach the learner to check whether the expression works figuratively and what it communicates before declaring it nonsensical.
31. Context can make unusual sentences meaningful
A science-fiction story may describe a robot tasting sunlight. That can be coherent within the invented world even if the words are unusual in everyday descriptions.
A researcher must control what participants know about the situation. In teaching, the student should distinguish fictional premises from real-world assertions and interpret the sentence according to its discourse context.
32. A false fact can sound familiar
A commonly repeated false claim may use highly familiar words and a plausible grammatical pattern. It could be easy to process while remaining incorrect.
Do not confuse neural ease of semantic access with independent truth verification. In History, Science and General Paper, students must consult reliable evidence rather than regard fluency as proof. Vocabulary helps organise claims, not certify them.
33. Clinical research has a different purpose
N400 has been studied in research involving clinical populations and language disorders. That work requires specialised designs, participant selection and professional interpretation.
A teacher cannot diagnose dyslexia, autism or intelligence from a pupil’s reaction to one surprising sentence. Such an inference would exceed the evidence. Classroom assessment should focus on comprehension and refer persistent concerns to appropriate professionals.
34. Ethical EEG research needs informed participation
Recording EEG from participants requires appropriate consent, comfortable procedures and privacy protection. Research reports should describe relevant recruitment and exclusions, especially when working with children.
A public vocabulary guide can explain the phenomenon through original written examples without collecting brain data. The value lies in methodological understanding rather than encouraging unnecessary or unvalidated monitoring of pupils.
35. A high-school explanation can use two sentences
Show a natural sentence ending and an unexpected but grammatical ending. Ask students to identify which is more supported by the context and why. Explain that a real experiment could compare EEG responses to such controlled conditions.
The class is not measuring an N400. It is learning what variable the researcher might change and what alternative explanations need consideration. That distinction is part of scientific literacy.
36. Primary English: identify the contextual clue
A Primary pupil might see ‘The gardener watered the …’ and compare flowers with an unlikely object. The teacher asks what words in the sentence guide the expectation.
The pupil then creates an original sentence using the less familiar noun appropriately. The exercise builds predictive reading and flexible meaning without teaching that unusual objects are always impossible. Context is a clue, not a prison.
37. Secondary English: avoid automatic substitutions
A student may replace a common accurate word with a rare near-synonym merely for vocabulary variety. The new form may fit poorly with the surrounding grammar or meaning.
Ask the learner to state the intended proposition and evaluate whether the replacement preserves it. The N400 concept motivates attention to context, but a teacher should not claim to observe the child’s brain response from the sentence alone.
38. Science: a surprising fact still deserves testing
A science passage can introduce an unfamiliar mechanism that contradicts a student’s prior assumption. An unexpected term may demand explanation yet be correct.
The next step is to test the statement against the concept, experiment or scientific reference. The learner should not reject a fact because it violates expectation. Accurate vocabulary must connect to verified physical mechanisms.
39. Mathematics: language controls the operation
‘At least five’ includes five, whereas ‘more than five’ does not. These expressions differ by short, familiar words that change the set of permitted values.
The student should translate the phrase into an inequality and test a boundary case. N400 research does not solve the calculation; it illustrates why context and conceptual expectation matter during comprehension.
40. History: familiar narrative can obscure evidence
A historical claim may sound predictable because it resembles a familiar story. Yet the true record can contain unexpected dates, actions and motives.
Teach students to separate what the narrative leads them to anticipate from what sources actually support. This avoids treating a plausible-sounding phrase as confirmation of historical accuracy.
41. General Paper: coherence is not enough
A polished argument can contain coherent phrases while lacking reliable evidence. An essay can also present an initially surprising finding that becomes persuasive once the source and causal explanation are clear.
The student should evaluate definitions, warrants, source quality and counterarguments. N400 is a measure of some aspects of processing, not a substitute for critical reasoning.
42. Translation: preserve the intended sense
A translator can choose a common target-language word that sounds plausible but changes a prohibition, agency or technical relationship. The sentence may become easy to process while meaning the wrong thing.
Check the source proposition and the target-language context. Accuracy of translation requires preserving relations and conditions, not simply producing familiar collocations.
43. A reproducible research notebook
A good note lists the target word, sentence context, alternative completion, cloze norms if used, participant language, modality, electrode and analysis choices, and the observed condition difference.
Without these details, a claim about a ‘brain signal for wrong words’ is too broad. A learner who can state what was actually compared has gained a general research skill.
44. The final lesson: meaning is constructed in context
N400 research provides a powerful indication that word processing changes with what came before and how concepts relate. It helps connect vocabulary knowledge to measurable, time-sensitive comprehension.
But the educational goal goes beyond noticing unexpected words. A learner must identify the right sense, question unsupported claims and use new expressions accurately. Vocabulary is technology when it supports that reliable chain from word to understanding.
Full case: prediction versus comprehension
A fictional learner completes ‘The gardener watered the …’ with flowers. A second learner suggests seedlings. Both may be appropriate in a garden. The teacher then gives an unfamiliar technical object that also might be watered and explains the context in which it belongs. A low initial expectation is not a reason to reject a valid expression.
The pupils identify which words gave the clue, state what additional information changed the interpretation and write accurate new sentences. The assessment is based on explanation and transfer, not a claim about the children’s brain signals.
Eight-step vocabulary context audit
- Identify the target word and its exact sense in the passage.
- State what the preceding words make likely and why.
- Distinguish semantic support from grammatical well-formedness.
- Check whether an unusual completion could make sense with additional context.
- Consult a dictionary or authoritative subject source when the term is unfamiliar.
- Ask what evidence makes a factual claim true rather than merely plausible.
- Rewrite the idea for a fresh audience without changing its meaning.
- Evaluate comprehension and independent use instead of speed or assumed neural responses.
Frequently asked questions
1. What is the N400?
An ERP response associated with meaning processing, often observed as a negative-going difference around the 400-ms range.
2. Who discovered its famous language effect?
Marta Kutas and Steven Hillyard published the landmark semantic-incongruity result in 1980.
3. Does N400 mean every sentence is wrong?
No. It is sensitive to semantic support and other factors, not a universal lie detector.
4. Is 400 milliseconds the exact time everyone understands a word?
No. Component timing varies, and interpretation is not a single instantaneous event.
5. How is it measured?
Typically through EEG and averaged event-related responses under controlled comparisons.
6. Does a larger N400 imply lower intelligence?
No. It is a condition-dependent experimental measure, not a personal intelligence score.
7. How does cloze probability relate?
Completion norms estimate how predictable a target word is for sampled participants.
8. Can pictures elicit relevant N400 effects?
Yes. The response is studied with meaningful nonlinguistic stimuli as well as words.
9. Is N400 exactly the opposite of P600?
No. Their polarity and typical associations differ, but the underlying language processes are not a simple two-error classification.
10. What is the useful educational insight?
Words acquire specific meaning through context, and expectation must be checked against evidence.
Primary research and linked vocabulary learning
- Kutas and Hillyard (1980): Reading Senseless Sentences — Brain Potentials Reflect Semantic Incongruity — the landmark Science report.
- Kutas and Hillyard (1980): Event-Related Brain Potentials to Semantically Inappropriate and Surprisingly Large Words — contrasts semantic and visual deviations.
- Kutas and Hillyard (1984): Brain Potentials During Reading Reflect Word Expectancy and Semantic Association — N400 and cloze expectancy.
- Kutas and Federmeier (2011): Thirty Years and Counting — Finding Meaning in the N400 — major scholarly review of stimulus types, meaning and theory.
- Review: Functional Roles of N400 and P600 in Semantic Anomalies — explanation of limitations of simple meaning-versus-grammar labels.
- Vocabulary | What Is a Word-Sense Inventory? — distinguish senses in context.
- Vocabulary | What Is a Semantic Relatedness Judgement? — connection versus synonymy.
- Vocabulary | What Is Masked Priming? — another laboratory method rather than a synonym for N400.
- Vocabulary Learning Hub — continue learning how word knowledge supports thought.
N400 research makes visible that meaning is processed in context, not merely looked up after a sentence ends. A strong vocabulary system helps a learner use that context without mistaking a familiar expectation for an authoritative fact. Accurate reading is the bridge from prediction to knowledge.