恒挚 Share | Event-related potentials (ERPs): an in-depth analysis from neural origins to cognitive components

NEURAL ORIGIN

Simply put, ERPs are the specific electrical signals that the brain produces in response to a particular "event" (e.g., seeing a word, hearing a sound, performing an action). These signals are hidden in the clutter of the electroencephalogram (EEG), and need to be "extracted" by technical means in order to become the key to deciphering brain activity.

Its best ability is "time resolution" - can be accurate to the millisecond level. For example, if you see the word "apple", it takes 100 milliseconds for the brain to recognize it as a word, and 300 milliseconds to associate it with its flavor, and the ERP can clearly record these subtle time differences. Even fMRI (Functional Magnetic Resonance Imaging) can't match this (fMRI has a time resolution of seconds).

How do ERPs "pick up" brain signals?

In fact, the ERP signal triggered by a single event is so weak (only a few microvolts, equivalent to one ten-millionth of a volt) that it can easily be drowned out by the "noise" of spontaneous brain electrical activity, muscle flutter, heartbeat, and so on. It's like trying to hear someone in a noisy food market - what do you do?

The answer is "superimposed averaging": the same event (e.g., looking at the same picture over and over again) is repeated dozens or even hundreds of times, and the signals recorded each time are superimposed. In this way, the random "noise" will cancel each other out, while the "effective signal" associated with the event will grow and eventually become clear.

ERP signals are manifested as a series of positive and negative waveforms on the scalp, which are usually named after "polarity (N is negative, P is positive) + latency (milliseconds)" (e.g., N100, P300). Based on latency and cognitive function, ERP components can be categorized into exogenous (early component), which is mainly related to the physical properties of the stimulus, and endogenous (middle and late component), which reflects higher cognitive processing.

The exogenous component (also known as the early component) usually appears within 100 ms of stimulus presentation, is highly sensitive to the physical characteristics of the stimulus (e.g., intensity, frequency, and location), is not significantly affected by the subjective state of the individual (e.g., attention, motivation), and mainly reflects the early processing of the sensory system.

Brainstem evoked potentials (BAEP/ABR):

With a latency of about 1-10 ms, it originates from auditory nuclei in the brainstem (e.g., cochlear nucleus, superior olivary nucleus) and appears only in response to auditory stimulation. Its waveform is stable (e.g., I-V wave) and is commonly used to assess the integrity of the auditory pathway (e.g., newborn hearing screening, diagnosis of acoustic neuroma). For example, prolonged V-wave latency in brainstem evoked potentials may indicate auditory nerve damage.

Visual P1 and N1 (early visual components):

P1: latency about 80-120ms, mainly distributed in occipital cortex (visual area), wave amplitude rises with the increase of brightness and contrast of the stimulus, reflecting the early processing from the retina to the primary visual cortex (area V1).

N1: latency ~150-200ms, more widely distributed than P1, associated with feature recognition (e.g., direction, color) of visual stimuli. It has been found that the amplitudes of P1 and N1 are significantly enhanced when the stimulus appears at the location where attention is directed - suggesting that early sensory processing has been modulated by attention.

Auditory N1 (early auditory component): with a latency of about 100-150 ms, it is distributed in the temporal cortex (auditory area) and is sensitive to changes in the frequency and intensity of sounds. For example, the sudden appearance of a sound induces a larger N1 amplitude, while the repeated presentation of a sound leads to a gradual weakening of the N1 (i.e., the phenomenon of "habituation"), reflecting the auditory system's rapid detection of novel stimuli.

The mid-latency component, which occurs 100-300 ms after stimulus presentation, is influenced by both the physical properties of the stimulus and primary cognitive processes (e.g., stimulus recognition, conflict detection), and serves as a "bridge" connecting sensory input to higher cognition.

N200 (N2): latency around 200ms, distribution varies depending on the task (e.g., frontal area, central area), core cognitive significance is related to "conflict monitoring" and "stimulus novelty detection".

In the "oddball paradigm" (where deviant stimuli are occasionally presented that differ from the standard stimulus), deviant stimuli evoke N2 in frontal areas, reflecting the brain's early recognition of abnormal stimuli.

In the Stroop task (e.g., writing the word "green" in red), semantic conflict enhances N2 amplitude, suggesting its involvement in cognitive conflict detection and resolution.

In addition, N2 is associated with response inhibition: in the Go/No-Go task, the amplitude of N2 in the central region was significantly elevated when it was necessary to inhibit a dominant response (e.g., not pressing a key to a No-Go stimulus).

P200 (P2): latency of ~150-250ms, widely distributed across the scalp, associated with feature integration of stimuli (e.g., combining color and shape information from visual stimuli) and attentional allocation. For example, when the stimulus is related to a task goal, the P2 amplitude increases, reflecting a skewing of attentional resources toward the relevant stimulus.

The late latency component appears more than 300ms after stimulus presentation, is almost unaffected by the physical properties of the stimulus, and is mainly related to higher cognitive processes such as attention, memory, language, and decision-making, and is the component that has received the most attention in ERP studies.

P300 (P3): as one of the most widely studied components of the ERP, P300 has a latency of about 300-600ms, is mainly distributed in the parietal and central regions, and can be subdivided into two subcomponents, P3a and P3b:

P3a: slightly earlier latency (~300-400ms), distributed in frontal areas, sensitive to novel stimuli (e.g., sudden appearance of extraneous sounds), reflecting shifts in automatic attention ("facing response").

P3b: The latency is slightly later (~350-600ms), distributed in parietal regions, and closely related to "working memory updating" and "task-relevant processing". In the oddball paradigm, deviant stimuli (task-relevant) evoke P3b, the amplitude of which is positively correlated with the unexpectedness of the stimulus (the more unexpected, the larger the amplitude), and the latency is correlated with the speed of information processing (the faster the processing, the shorter the latency).

Clinical studies have found that patients with Alzheimer's disease have prolonged P3b latency and reduced wave amplitude, suggesting a decrease in cognitive processing speed and memory updating, while patients with schizophrenia have abnormal P3b wave amplitude, which may reflect deficits in attention and information integration functions.

N400: Named after the negative wave that occurs at about 400ms, mainly in the central-parietal and temporal regions, it is a core indicator of language processing research, especially related to "semantic integration".

In sentence comprehension, when a sentence-final word semantically conflicts with the preceding text (e.g., "He added salt to his coffee"), the amplitude of the N400 is significantly larger than in semantically coherent sentences (e.g., "He added sugar to his coffee"), and the more severe the conflict, the larger the amplitude.

In addition to language, the N400 is also involved in non-linguistic semantic processing, such as in the picture-word matching task, when pictures such as("cat") also induced the N400 when it did not match the word ("dog"), suggesting that it reflects more general "inter-conceptual correlation" processing.

Research in developmental psychology has shown that infants can produce N400-like responses to semantic conflict at around 18 months of age, suggesting that human semantic processing skills germinate early in language development.

Late Positive Component (LPC): latency about 500-1000ms, widely distributed (parietal area predominant), closely related to memory extraction and emotional processing.

In the rerecognition memory task, the LPC amplitude of the "seen stimulus" (old stimulus) was significantly larger than that of the "unseen stimulus" (new stimulus), and this "old-new effect" This "old-new effect" reflects the successful completion of memory extraction, and the larger the amplitude, the higher the memory clarity.

In emotional stimulus processing, negative emotional stimuli (e.g., fearful faces) evoked greater LPC amplitudes than neutral stimuli, indicating their involvement in the deep processing and evaluation of emotional information.

Error-Related Negative Wave (ERN/Ne): With a latency of about 50-100ms (after an error response), ERN is distributed in the central frontal area and is a core indicator of "behavioral monitoring". When an individual realizes that he or she has made a wrong response (e.g., keystroke error), the amplitude of ERN will rise significantly, reflecting the rapid detection of behavioral errors in the anterior cingulate cortex. It has been found that anxious individuals have greater ERN amplitude, indicating greater sensitivity to errors, while ADHD patients have reduced ERN amplitude, which may be associated with deficits in their behavioral monitoring abilities.

It should be emphasized that the correspondence between ERP components and cognitive processes is not a simple "one-to-one" relationship, but rather a double complexity:

One component is multifunctional: e.g., N2 is involved in both conflict monitoring and is associated with response inhibition; P300 reflects both memory updating and attention allocation. This multifunctionality stems from the overlapping nature of cognitive processes - the same neural circuit may be involved in multiple tasks.

Multi-component: For example, "semantic processing" not only involves N400, but also may activate P2 (early feature extraction) and LPC (deep integration); "attentional modulation" may affect multiple components, such as P1, N1, P2, and so on.

Therefore, interpreting ERP data needs to be done in the context of the specific experimental design (e.g., task type, stimulus characteristics), and confounders need to be eliminated through the control variables approach in order to more accurately infer cognitive meaning.

From the perspective of neural origin, ERP is the macroscopic embodiment of the synchronized activity of cortical pyramidal cells, and it is the electrical signal imprint of the "collective wisdom" of neuronal groups; from the perspective of cognitive components, from the early P1, N1, to the late P300, N400, each waveform is the key to unlocking the riddle of perception, attention, language, and memory. Despite the limitation of spatial resolution, ERP, with its millisecond-level temporal accuracy, is still the "gold standard" for studying the dynamic mechanism of cognitive processes.

Nowadays, ERP has been widely used in basic research (e.g., neural mechanisms of infant language development and decision-making), clinical diagnosis (e.g., early screening of epilepsy and autism), and real life (e.g., the "P300 speller" in the Brain-Machine-Interface to help people with disabilities to communicate). With technological advances (e.g., high-density electrode caps, advanced signal processing algorithms), ERP will continue to play an irreplaceable role in the journey of "reading the brain", allowing us to better understand the nature of the human mind - the neural activities that occur in milliseconds. The neural activity that occurs in milliseconds is the source of thought, emotion and intelligence.