is a component of time-locked EEG
signals known as event-related potentials
(ERP). It is a negative-going deflection that peaks around 400 milliseconds post-stimulus onset, although it can extend from 250-500 ms, and is typically maximal over centro-parietal electrode
sites. The N400 is part of the normal brain
response to words
and other meaningful (or potentially meaningful) stimuli, including visual and auditory words, sign language
, environmental sounds, and smells.
The N400 was first discovered by Marta Kutas
and Steven Hillyard in 1980.
They conducted the first experiment looking at the response to unexpected words in read sentences
, expecting to elicit a P300
component. The P300 had previously been shown to be elicited by unexpected stimuli. Kutas and Hillyard therefore used sentences with anomalous endings (i.e.I take coffee with cream and dog
), expecting to see a P300 to the unexpected sentence-final words. However, instead of eliciting a large positivity, these anomalous endings elicited a large negativity, relative to the sentences with expected endings (i.e. He returned the book to the library
) In the same paper, they confirmed that the negativity was not just caused by any unexpected event at the end of a sentence, since a semantically expected but physically unexpected word (i.e. She put on her high-heeled SHOES
) elicited a P300 instead of negativity in the N400 window. This finding showed that the N400 is related to semantic processing, and is not just a response to unexpected words.
The N400 is characterized by a distinct pattern of electrical activity that can be observed at the scalp. As its name indicates, this waveform peaks around 400 ms post-stimulus onset, with negativity that can be observed in the time window ranging from 250-500 ms. This latency (delay between stimulus and response) is very stable across tasks—little else besides age
affects the timing of its peak.
The N400 is a negative component, relative to reference electrodes
placed on the mastoid
processes (the bony ridge behind the ear), and relative to a 100 ms pre-stimulus baseline. Its amplitude can range from -5 to 5 microvolts
. However, it is important to note that in studies using the N400 as a dependent measure, the relative amplitude
of the waveform compared to another experimental condition (the "N400 effect") is more important than its absolute amplitude. The N400 itself is not always negative—it is just a more negative-going deflection than that seen to other conditions. Its distribution is maximal over centro-parietal electrode sites, and is slightly larger over the left side of the head for visual words, although the distribution can change slightly depending on the eliciting stimulus.
A typical experiment designed to study the N400 will usually involve the visual presentation of words, either in sentence or list contexts. In a typical visual N400 experiment, for example, subjects will be seated in front of a computer monitor
while words are presented one-by-one at a central screen location. Stimuli must be presented centrally because eye movements
will generate large amounts of electrical noise that will mask the relatively small N400 component. Subjects will often be given a behavioral task (e.g., making a word/nonword decision, answering a comprehension question, responding to a memory
probe), either after each stimulus or at longer intervals, to ensure that subjects are paying attention. Note, however, that overt responses by the subject are not required to elicit the N400—passively viewing stimuli will still evoke this response.
An example of an experimental task used to study the N400 is a priming
paradigm. Subjects are shown a list of words in which a prime word is either associatively related to a target word (e.g. bee and honey), semantically related (e.g. sugar and honey) or a direct repetition (e.g. honey and honey). The N400 amplitude seen to the target word (honey) will be reduced upon repetition due to semantic priming.
The amount of reduction in amplitude can be used to measure the degree of relatedness between the words.
Another widely used experimental task used to study the N400 is sentence reading
. In this kind of study, sentences are presented to subjects centrally, one word at a time, until the sentence is completed. Alternatively, subjects could listen to a sentence as natural auditory speech. Again, subjects may be asked to respond to comprehension questions periodically throughout the experiment, although this is not necessary. Experimenters can choose to manipulate various linguistic characteristics of the sentences, including contextual constraint or the cloze probability of the sentence-final word (see below for a definition of cloze probability) to observe how these changes affect the waveform's amplitude.
As previously mentioned, the N400 response is seen to all meaningful, or potentially meaningful, stimuli. As such, a wide range of paradigms can be used to study it. Experiments involving the presentation of spoken words,
pictures embedded at the end of sentences,
words related to current context or orientation
and videos of real-word events,
have all been used to study the N400, just to name a few.
Extensive research has been carried out to better understand what kinds of experimental manipulations do and do not affect the N400. General findings are discussed below.
Factors that affect N400 amplitude
The frequency of a word's usage is known to affect the amplitude of the N400. With all else being constant, highly frequent words elicit reduced N400s relative to infrequent words.
As previously mentioned, N400 amplitude is also reduced by repetition, such that a word's second presentation exhibits a more positive response when repeated in context.
These findings suggest that when a word is highly frequent or has recently appeared in context, it eases the semantic processing thought to be indexed by the N400, reducing its amplitude.
N400 amplitude is also sensitive to a word's orthographic neighborhood size, or how many other words differ from it by only one letter
). Words with large neighborhoods (that have many other physically similar items) elicit larger N400 amplitudes than do words with small neighborhoods.
This finding also holds true for pseudowords, or pronounceable letter strings that are not real words (e.g. flom), which are not themselves meaningful but look like words. This has been taken as evidence that the N400 reflects general activation in the comprehension
network, such that items that look like many words (regardless of whether it itself is a word) partially activate the representations of similar-looking words, producing a more negative N400.
The N400 is sensitive to priming
: in other words, its amplitude is reduced when a target word is preceded by a word that is semantically, morphologically, or orthographically related to it.
In a sentence context, an important determinant of N400 amplitude elicited by a word is its cloze probability. Cloze probability is defined as the probability of the target word completing that particular sentence frame. Kutas and Hillyard (1984) found that a word's N400 amplitude has a nearly inverse linear relationship with its cloze probability.
That is, as a word becomes less expected in context, its N400 amplitude is increased relative to more expected words. Words that are incongruent with a context (and thus have a cloze probability of 0) elicit large N400 amplitudes as well (although the amplitude of the N400 for incongruent words is also modulated by the cloze probability of the congruent word that would have been expected in its place
Relatedly, the N400 amplitude elicited by open-class words (i.e. nouns, verbs, adjectives, and adverbs) is reduced for words appearing later in a sentence compared to earlier words.
Taken together, these findings suggest that when the prior context builds up meaning, it makes the processing of upcoming words that fit with that context easier, reducing the N400 amplitude they elicit.
Factors that do not affect N400 amplitude
While the N400 is larger to unexpected items at the end of a sentence, its amplitude is generally unaffected by negation
that causes the last word to be untrue and thus anomalous.
For example, in the sentence A sparrow is a building
, the N400 response to building
is more negative than the N400 response to bird in the sentence A sparrow is a bird
. In this case, building
has a lower cloze probability, and so it is less expected than bird
. However, if negation
is added to both sentences in the form of the word not
(i.e. A sparrow is not a building
and A sparrow is not a bird
), the N400 amplitude to building
will still be more negative than that seen to bird
. This suggests that the N400 responds to the relationship between words in context, but is not necessarily sensitive to the sentence's truth value
. More recent research, however, has demonstrated that the N400 can sometimes be modulated by quantifiers or adjectives that serve negation-like purposes,
or by pragmatically
Additionally, grammatical violations do not elicit a large N400 response. Rather, these types of violations show a large positivity from about 500-1000 ms after stimulus onset, known as the P600
Factors that affect N400 latency
A striking feature of the N400 is the general invariance of its peak latency. Although many different experimental manipulations affect the amplitude of the N400, few factors (aging and disease states and language proficiency being rare examples) alter the time it takes for the N400 component to reach a peak amplitude.
Although localization of the neural generators of an ERP signal is difficult due to the spreading of current from the source to the sensors, multiple techniques can be used to provide converging evidence about possible neural sources.
Using methods such as recordings directly off the surface of the brain or from electrodes implanted in the brain, evidence from brain damaged patients, and magnetoencephalographic (MEG)
recordings (which measure magnetic activity at the scalp associated with the electrical signal measured by ERPs), the left temporal lobe
has been highlighted as an important source for the N400, with additional contributions from the right temporal lobe
More generally, however, activity in a wide network of brain areas is elicited in the N400 time window, suggesting a highly distributed neural source.
There is still much debate as to exactly what kind of neural and comprehension processes the N400 indexes. Some researchers believe that the underlying processes reflected in the N400 occur after a stimulus has been recognized. For example, Brown and Hagoort (1993) believe that the N400 occurs late in the processing stream, and reflects the integration of a word's meaning into the preceding context (see Kutas & Federmeier, in press,
for a discussion). However, this account has not explained why items that themselves have no meaning (e.g. pseudowords
without defined associations) also elicit the N400. Other researchers believe that the N400 occurs much earlier, before words are recognized, and represents orthographic
More recent accounts posit that the N400 represents a broader range of processes indexing access to semantic memory
. According to this account, it represents the binding of information obtained from stimulus input with representations from short- and long-term memory
(such as recent context, and accessing a word's meaning in long term memory) that work together to create meaning from the information available in the current context (Federmeier & Laszlo, 2009; see Kutas & Federmeier, in press
Another account is that the N400 reflects prediction error or surprisal. Word-based surprisal was a strong predictor of N400 amplitude in an ERP corpus.
In addition, connectionist models make use of prediction error for learning and linguistic adaptation, and these models can explain several N400/P600 results in terms of prediction error propagation for learning.
As research in the field of electrophysiology continues to progress, these theories will likely be refined to include a complete account of just what the N400 represents.
- ^ a b c Kutas, M.; Federmeier, K.D. (2000). "Electrophysiology reveals semantic memory use in language comprehension". Trends in Cognitive Sciences. 4 (12): 463–470. doi:10.1016/s1364-6613(00)01560-6. PMID 11115760.
- ^ a b c d e f g Kutas, Marta; Federmeier, Kara D. (2011). "Thirty Years and Counting: Finding Meaning in the N400 Component of the Event-Related Brain Potential (ERP)". Annual Review of Psychology. 62: 621–647. doi:10.1146/annurev.psych.093008.131123. PMC 4052444.
- ^ (See Kutas & Federmeier, 2009, for review)
- ^ Kutas, M.; Hillyard, S. A. (1980). "Reading senseless sentences: Brain potentials reflect semantic incongruity". Science. 207 (4427): 203–208. doi:10.1126/science.7350657. PMID 7350657.
- ^ Van Petten, C.; Coulson, S.; Rubin, S.; Plante, E.; Parks, M. (1999). "Time course of word identification and semantic integration in spoken language". Journal of Experimental Psychology: Learning, Memory, and Cognition. 25 (2): 394–417. doi:10.1037/0278-7322.214.171.1244.
- ^ Laszlo, S.; Federmeier, K.D. (2008). "Minding the PS, queues, and PXQs: Uniformity of semantic processing across stimulus types". Psychophysiology. 45 (3): 458–466. doi:10.1111/j.1469-8986.2007.00636.x. PMC 2704151. PMID 18221447.
- ^ Federmeier, K.D.; Kutas, M. (2001). "Meaning and modality: Influences of context, semantic memory organization, and perceptual predictability on picture processing". Journal of Experimental Psychology: Learning, Memory, and Cognition. 27 (1): 202–224. CiteSeerX 10.1.1.423.6663. doi:10.1037/0278-73126.96.36.199.
- ^ Daltrozzo, J.; Schön, D. (2009). "Conceptual processing in music as revealed by N400 effects on words and musical targets" (PDF). Journal of Cognitive Neuroscience. 21 (10): 1882–1892. CiteSeerX 10.1.1.372.401. doi:10.1162/jocn.2009.21113. PMID 18823240.
- ^ Ghosh Hajra, S.; Liu, C.C.; Song, X.; Fickling, S.; Cheung, T.P.L.; D’Arcy, R. (2018). "Accessing knowledge of the 'here and now': a new technique for capturing electromagnetic markers of orientation processing". Journal of Neural Engineering. 16 (1): 016008. doi:10.1088/1741-2552/aae91e. PMID 30507557.
- ^ Sitnikova, T.; Kuperberg, G.; Holcomb, P.J. (2003). "Semantic integration in videos of real0world events: an electrophysiological investigation". Psychophysiology. 40: 160–164. doi:10.1111/1469-8986.00016. PMID 12751813.
- ^ a b Van Petten, C.; Kutas, M. (1990). "Interactions between sentence context and word frequency in event-related brain potentials". Memory and Cognition. 18 (4): 380–393. doi:10.3758/bf03197127.
- ^ Van Petten, C.; Kutas, M.; Kluender, R.; Mitchiner, M.; McIsaac, H. (1991). "Fractionating the word repetition effect with event-related potentials". Journal of Cognitive Neuroscience. 3 (2): 131–150. doi:10.1162/jocn.19188.8.131.52. PMID 23972089.
- ^ Holcomb, P.J.; Grainger, J.; O'Rourke, T. (2002). "An electrophysiology study of the effects of orthographic neighborhood size on printed word perception". Journal of Cognitive Neuroscience. 14 (6): 938–950. CiteSeerX 10.1.1.466.2444. doi:10.1162/089892902760191153. PMID 12191460.
- ^ Kutas, M.; Hillyard, S. A. (1984). "Brain potentials during reading reflect word expectancy and semantic association". Nature. 307 (5947): 161–163. doi:10.1038/307161a0.
- ^ Borovsky, A.; Kutas, M.; Elman, J. (2011). "Learning to use words: Event related potentials index single-shot contextual word learning". Cognition. 116 (2): 289–296. doi:10.1016/j.cognition.2010.05.004. PMC 2904319. PMID 20621846.
- ^ Fischler, I.; Bloom, P.A.; Childers, D.G.; Roucos, S.E.; Perry, N.W. (1983). "Brain potentials related to stages of sentence verification". Psychophysiology. 20 (4): 400–410. doi:10.1111/j.1469-8986.1983.tb00920.x. PMID 6356204.
- ^ Urbach, Thomas; Kutas, Marta (2010). "Quantifiers more or less quantify online: ERP evidence for partial incremental interpretation". Journal of Memory and Language. 63 (2): 158–79. doi:10.1016/j.jml.2010.03.008. PMC 2902883. PMID 20640044.
- ^ Nieuwland, Mante; Kuperberg, Gina (2008). "When the truth is not too hard to handle: an event-related potential study on the pragmatics of negation". Psychological Science. 19 (12): 1213–8. doi:10.1111/j.1467-9280.2008.02226.x. PMC 3225068. PMID 19121125.
- ^ Federmeier, K. D. and Laszlo, S. (2009). Time for meaning: Electrophysiology provides insights into the dynamics of representation and processing in semantic memory. In B. H. Ross (Ed.), Psychology of Learning and Motivation, Volume 51 (pp 1-44). Burlington: Academic Press.
- ^ Haan, H., Streb, J., Bien, S., & Ro, F. (2000). Reconstructions of the Semantic N400 Effect : Using a Generalized Minimum Norm Model with Different Constraints ( L1 and L2 Norm ), 192, 178–192.
- ^ Van Petten, C.; Luka, B. (2006). "Neural localization of semantic context effects in electromagnetic and hemodynamic studies". Brain and Language. 97 (3): 279–293. doi:10.1016/j.bandl.2005.11.003. PMID 16343606.
- ^ Deacon, D.; Dynowska, A.; Ritter, W.; Grose-Fifer, J. (2004). "Repetition and semantic priming of nonwords: Implications for theories of N400 and word recognition". Psychophysiology. 41 (1): 60–74. doi:10.1111/1469-8986.00120.
- ^ Frank, Stefan L.; Otten, Leun J.; Galli, Giulia; Vigliocco, Gabriella (2015). "The ERP response to the amount of information conveyed by words in sentences". Brain and Language. 140: 1–11. doi:10.1016/j.bandl.2014.10.006.
- ^ Fitz, Hartmut; Chang, Franklin (2019). "Language ERPs reflect learning through prediction error propagation". Cognitive Psychology. 111: 15–52. doi:10.1016/j.cogpsych.2019.03.002. hdl:21.11116/0000-0003-474F-6. PMID 30921626.
Last edited on 7 December 2020, at 22:51
Content is available under CC BY-SA 3.0
unless otherwise noted.