Decrement of the N1 auditory event-related potential with stimulus repetition: habituation vs. refractoriness

https://doi.org/10.1016/S0167-8760(98)00040-3Get rights and content

Abstract

We examined whether the amplitude decrement traditionally found for the N1 peak of the event-related potential (ERP) with repetition of auditory stimuli results from the process of habituation or from the refractory period of the neural elements underlying the N1 response. These competing accounts of the process underlying the N1 decrement with repetition differ in terms of the predicted effects of variations in stimulus repetition and interstimulus interval (ISI). These predictions were examined using a short-term habituation design with a factorial combination of stimulus repetition and ISI. Forty-five subjects received 21 stimulus trains, each consisting of seven innocuous tones, all at 1 kHz except the sixth, which was a 1.5-kHz tone. Each subject was assigned to one of three ISI conditions (either 1, 3 or 10 s). The results provide little support for the view that N1 response decrement with stimulus repetition reflects a process of habituation. The present results provide greater support for the view that this decrement is based on the separate refractory periods or recovery cycle processes of at least two neural generators contributing to activity in the N1 peak latency range.

Introduction

An important feature of the N1 peak of the auditory event-related potential is its systematic reduction in amplitude when the eliciting stimulus is repeated. A major psychophysiological issue regarding the functional nature of N1 amplitude decrement has been the extent to which this response decrement reflects a psychologically relevant process or a more basic neurophysiological process. Traditionally, this response decrement has been characterised either in terms of a process of habituation, such as that defined by Orienting Response theory (Sokolov, 1963), or a process involving the recovery cycle or refractory period of the neural generators underlying the N1 (Ritter et al., 1968; Callaway, 1973). One method of distinguishing between the distinct processes of habituation and refractoriness is that amplitude reductions due to refractoriness should stabilise immediately after repetition of a stimulus while habituation could entail a more progressive decline in responsiveness (Picton et al., 1976). If the N1 response decrement observed with stimulus repetition is the result of a refractory period or recovery cycle effect, then the primary determination of the decrement should be the interval between stimuli not repetition per se (Nelson and Lassman, 1968; Rothman et al., 1970; Roth et al., 1976).

According to Sokolov's Orienting Response theory, habituation reflects the establishment or updating of a neuronal model or template of a stimulus following repeated exposure. An orienting response results whenever a mismatch occurs between the established neuronal model and a new stimulus (Sokolov, 1963). In this context, auditory N1 response habituation refers to a decrement in response amplitude resulting from a loss of novelty associated with the building of a neuronal model following repetition of an auditory stimulus.

Recovery cycle or refractory period effects similarly predict response decrement following stimulus repetition. The recovery cycle concept in ERP research derives from studies of single nerve cells where there is a recovery cycle or refractory period following an action potential, during which time further stimulation will not result in an action potential. The assumption central to this view of the N1 response decrement is that polysynaptic neural systems or neural networks generating the N1 show a phenomenon similar to the excitability cycle of single cells. It is generally argued that the recovery period reflects the dissipation of a state of temporal excitability in the N1 generators following their activation by a stimulus (Callaway, 1973; Wastell, 1980). Although the nature of the gross neural mechanisms or networks underlying these refractory effects is not well understood (Näätänen, 1992), it is generally maintained that decrements are observed because of the temporal limitations inherent in the physiochemical mechanisms underlying N1 generation. Thus, closely spaced presentations of auditory stimuli do not allow adequate recovery of these mechanisms and a decline in N1 amplitude is observed (Callaway, 1973; Loveless, 1983; Näätänen and Picton, 1987).

Despite an extensive history of `short-term habituation' ERP studies (see Callaway, 1973; Loveless, 1983; Näätänen and Picton, 1987) investigating the functional significance of N1 response decrement, there remains a lack of consensus regarding the habituation/refractoriness issue. This relates, in part, to a tendency for some authors to use the term `habituation' merely to describe a rapid response decrement following repetition of an auditory stimulus. Others have cautioned against this loose usage, stating that response decrements can only be described as habituation when all other possible explanations, such as refractoriness, diminished arousal, sensory adaptation and receptor fatigue, have been ruled out (Thompson et al., 1973; Siddle et al., 1983; Barry et al., 1993). The common failure to define habituation adequately has led to substantial confusion when interpreting ERP response decrements, leading one author to describe the resulting literature as `recalcitrant' (Loveless, 1983, p.98). Thompson et al. (1973), assert that the common interpretation, that repeated stimulation results in habituation, is circular if the only defining feature is the response decrement itself, and that this `empty' definition of habituation `... represents no significant advancement over the assertion that it [response decrement] happens because God makes it happen' (p. 242).

Thompson and Spencer (1966)operational definition of habituation avoids such circularity and therefore allows a more exact determination as to whether N1 response decrements are consistent with habituation. Their criteria include a response decrement following stimulus repetition (as a negative exponential function of the number of stimulus repetitions), response recovery (an increase in responding to a change stimulus inserted in a train of repeated stimuli) and dishabituation (an increase in responding to a previously habituated stimulus following insertion of a change stimuli).

Although there is overwhelming evidence to show N1 amplitude decreases rapidly with stimulus repetition, (e.g. Ritter et al., 1968, Roeser and Price, 1969, Roth and Kopell, 1969, Weber, 1970, Fruhstorfer et al., 1970, Fruhstorfer, 1971, Ohman and Lader, 1972, Salamy and McKean, 1977, Prosser et al., 1981, Wood and Elmasian, 1986, Bourbon et al., 1987, Simons et al., 1987, Barry et al., 1992) the extent to which studies have produced the graded decrements following repeated stimulation presentation is less clear. These studies generally show asymptotic amplitude reductions of up to 50% by the second stimulus in the train, but here the results are somewhat inconsistent. Some studies have shown that N1 amplitude reductions do follow a negative exponential function of the number of stimulus presentations, and reach asymptote by the third or fourth stimulus in a train (e.g. Ritter et al., 1968; Fruhstorfer et al., 1970; Fruhstorfer, 1971), which is more consistent with the progressive response decrement characteristic of habituation.

While few ERP habituation studies have examined N1 response recovery to a stimulus change, evidence for this effect has been less consistent than that obtained for response decrement. An integration of this previous research is difficult as a result of differences in the design, stimulus parameters and the technique by which N1 activity is measured. Nevertheless, the conflicting results of several such studies indicate that N1 response recovery is not a stable phenomenon for short-term habituation designs (e.g. Ritter et al., 1968; Fruhstorfer, 1971; Salamy and McKean, 1977; Megela and Teyler, 1979; Wood and Elmasian, 1986; Barry et al., 1992).

Even fewer N1 habituation studies have incorporated a test for dishabituation in their design. Those that do, have found no evidence of N1 dishabituation (Fruhstorfer, 1971; Wood and Elmasian, 1986; Barry et al., 1992). Although some authors have reported dishabituation, as noted by Loveless (1983); Näätänen and Picton (1987) and Barry et al. (1992), this has often reflected the mis-interpretation of what constitutes dishabituation. It is now well established that dishabituation can only be identified as enhanced responding to a previously habituated stimulus following insertion of a changed stimulus, not enhanced responding to the changed stimulus itself (response recovery) (Barry et al., 1992). Barry et al. (1993)found that the electrodermal response showed dishabituation in the same paradigm that failed to demonstrate N1 dishabituation in Barry et al. (1992). This suggests that it is the N1 response, rather than the short-term habituation design, which fails to express this important criterion of habituation.

The sensitivity of the N1 peak to interstimulus interval (ISI) has been often exploited as a possible means of distinguishing between a decrement associated with neural refractoriness and those reflecting habituation (e.g. Ritter et al., 1968). Studies investigating the influence of ISI on N1 amplitude can be separated into those using short-term habituation designs and others generally described as `recovery cycle' studies (see Callaway, 1973and Loveless, 1983for reviews). Recovery cycle studies usually average ERP responses as a function of ISI in order to determine the temporal recovery function of the N1 response. This research has revealed a relatively stable inverse relationship between N1 amplitude and ISI (Davis, 1939Davis et al., 1966Zerlin and Davis, 1967Nelson and Lassman, 1968Webster, 1971Butler, 1973Nelson and Lassman, 1973Woods et al., 1980Woods and Courchesne, 1986). N1 amplitude is generally found to follow a linear function of the logarithm of ISI. Nelson and Lassman (1968)reported a 5.6 μV increase in N1 amplitude for every tenfold increase in ISI between 0.5 and 10 s. N1 amplitude generally increases dramatically for ISI increases from 0.5 s to two or 3 s and then increases more gradually, reaching a maximum at about 10 s. According to this literature N1 sensitivity to ISI is widely interpreted in terms of a physiological refractory period effect (Nelson and Lassman, 1968, Nelson and Lassman, 1973, Nelson and Lassman, 1977).

A clear distinction between response decrement associated with habituation and that associated with refractoriness is particularly difficult for ERP measures since amplitude reductions associated with neural refractoriness may be superimposed upon, or even interactive with, those resulting from habituation (Picton et al., 1976). Similarly, response recovery to a change in stimulation may be predicted by both processes, either as a result of the stimulus specificity of the neural elements underlying the N1 response (Butler, 1968, Butler, 1972; Näätänen et al., 1988) or due to the novelty associated with a change in stimulation (Sokolov, 1963). Despite the extensive N1 habituation literature, a formal examination of these issues has not been carried out. A better determination of the extent to which a decrement reflects habituation could be made if a thorough assessment of the established criteria of habituation is carried out, together with an examination of the degree to which this criteria are influenced by stimulus interval.

An important limitation of N1 habituation research has been the tendency to treat the N1 peak as a unitary cortical response. Following increasing interest in the neurogenesis of scalp-recorded electrical activity generally, N1 studies have produced an extensive body of evidence (including animal studies, lesion studies, multiple electrode EEG, magnetoencephalographic (MEG), intracerebral recordings and evoked response source analysis) which support the original proposition that multiple cerebral generators may contribute to the N1 peak (Vaughan and Ritter, 1970).

Initially, N1 sub-components were distinguished in terms of the scalp topography and latency of negative peaks in the N1 latency region. McCallum and Curry (1980)detailed three peaks: N1a (fronto-temporal electrodes/75 ms), N1b (central electrodes (vertex potential)/100 ms) and N1c (mid-temporal electrodes/130 ms). Wolpaw and Penry (1975)described a positive–negative peak sequence recorded at mid-temporal electrodes as a `T complex' and labelled the peaks Ta and Tb, respectively. Tb is equivalent to the N1c of McCallum and Curry (1980)and Ta can be identified as a small positive peak that separates N1a and N1c at mid-temporal electrodes.

More recently, distinctions between N1 sub-components are made in terms of neural generators or sources rather than the latency and topography of individual N1 peaks. This approach addresses the problem of superposition that most likely confounds a peak analysis when spatially and temporally overlapping sources contribute to a single ERP peak recorded at the scalp. Näätänen and Picton (1987) reviewed evidence supporting their classification of three different generators or `true' N1 components that contribute to the scalp recorded N1 peak(s). Component 1, which has a fronto-central scalp distribution, is thought to reflect the activity of bilateral vertically oriented dipoles on the supratemporal plane in the primary auditory cortex. Component 2 is observed at mid-temporal scalp locations and is thought to reflect bilateral radially oriented dipoles located in the auditory association cortex at the superior temporal gyrus. Component 3 has a widespread scalp distribution, being maximally recorded at central scalp locations, suggesting cortical projection of a diffuse non-specific system (Loveless and Hari, 1993), with a radially oriented dipole located in the frontal cortex (Näätänen and Picton, 1987; Näätänen, 1992). Woods (1995)details a more comprehensive classification scheme, based on an original distinction between midline and temporal N1 sources of Wolpaw and Penry (1975), distinguishing six N1 components on the basis of tonotopy, integration time, latency, refractory properties, source orientation and location. Additional N1 sources in the frontal cortex have also recently been proposed on the basis of current source density and dipole analysis (Giard et al., 1994) as well as frontal sources only apparent at long ISIs (Alcaini et al., 1994).

Initial evidence that N1 sub-components may subserve functional differences came from a unique study examining the relationship between evoked MEG and EEG activity in the N1 latency range (N100m and N100, respectively). In a combined MEG/EEG study Hari et al. (1982)demonstrated that two N1 components can be distinguished on the basis of their distinct refractory properties. The existence of two spatially and temporally distinct bilateral N1 sources (N100m and L100m) differentially sensitive to ISI was confirmed by Lu et al. (1992). Essentially similar results were reported by Sams et al. (1993), who also found that two sources were required to account for N100m sensitivity with ISIs of 0.75 and 12 s. A later anterior component (N100ma) showed a relatively long recovery period, increasing in strength for intervals up to at least 12 s and an earlier posterior component (N100mp) had a much shorter recovery period, reaching a plateau at 6 s intervals. The data of Sams et al. (1993)has since been largely replicated by Loveless et al. (1996)and McEvoy et al. (1997)who also distinguished two components with different refractory properties and source locations contributing to the N100m; a long refractory component (N100ma) that was more anterior (and more superior) and peaked later (22–30 ms) than an earlier and more posterior short refractory component (N100mp).

Näätänen, 1990, Näätänen, 1992) suggests that the large amplitude and vertex maximum N1 peak with a long refractory period, such as that commonly elicited by the first stimulus in short-term habituation designs, is primarily due to the `non-specific' N1 component (component 3). The rapid decrement typically observed across subsequent stimulus positions reflects both the absence of the non-specific N1 component and the refractoriness of the supratemporal component (i.e. component 1). This non-specific component is thought to reflect an attentional triggering mechanism (Näätänen and Picton, 1987; Näätänen, 1992) and has been widely interpreted as a potential ERP index of the orienting response. However, as discussed, extensive ERP studies of habituation have yet to provide evidence that the activity underlying the N1 peak expresses some essential criteria of habituation. These features of the N1 decrement with repetition would be expected if N1 component activity subserves the orienting response.

A striking exception to this pattern of results is a recent MEG study by Lu et al. (1992)that provides evidence for the involvement of a posterior N1 component with a refractory period similar to that of component 3. Lu et al. (1992)reported that the L100m expressed several of the essential habituation criteria defined by Thompson and Spencer (1966). Although no direct data were provided, Lu et al. (1992)refer to a short-term habituation-like experiment where their L100m component showed three important criteria of habituation, a response decrement, response recovery and dishabituation.

Therefore the aim of the present study was to further examine the functional significance of the N1 response decrement by assessing the separate influence of ISI and stimulus repetition in the context of the traditional short-term habituation design. Important information concerning the nature of the process underlying this decrement could be obtained if some established criteria of habituation were assessed according to the degree to which these effects are influenced by stimulus interval and/or repetition. Further information regarding the functional significance of the N1 sub-components could also be obtained if differences in the latency and topographic distribution of the N1 decrement following stimulus repetition were examined.

Section snippets

Subjects and procedure

Forty-five healthy subjects (22 females and 23 males) aged 22–46 years participated in the experiment. Subjects were seated comfortably in a reclined position in a sound attenuated room and were instructed to keep their eyes fixated on a red coloured circle 1 cm in diameter, subtending 2° in visual space and displayed on a computer monitor placed 60 cm in front of the subject. Subjects were instructed to avoid, as much as possible, any eye movements, blinks or other muscle activity and to

Results

Despite near-perfect performance by most subjects on the distractor task (7.0±0.3 colour changes reported) and instructions allowing subjects to ignore the auditory stimuli, all subjects were aware of the tones in the headphones and several reported, without prompting, that the tones seemed to have been presented in a regular sequence. Although this limits the extent to which the subjects can be described as ignoring the auditory stimuli, their attention to these stimuli can be described as

Discussion

The present results were generally consistent with the view that the N1 amplitude decrement following stimulus repetition reflects a refractory process rather than habituation as the decrement was primarily determined by the interval between stimuli. Stimulus repetition, when considered separately from or in addition to the influence of stimulus interval, provided no significant effect on N1 decrement. Increasing ISI beyond the interval where an initial response decrement was observed did not

Acknowledgements

This research was supported by an Australian Postgraduate Award with Stipend to T.KB., and the Australian National Health and Medical Research Council.

References (64)

  • R Näätänen et al.

    Frequency and location specificity of the human vertex N1 wave

    Electroencephalogr. Clin. Neurophysiol.

    (1988)
  • A Ohman et al.

    Selective attention and habituation of the auditory averaged evoked response in humans

    Physiol. Behav.

    (1972)
  • N Perrault et al.

    Event-related potentials recorded from the scalp and nasopharynx. I. N1 and P2

    Electroencephalogr. Clin. Neurophysiol.

    (1984)
  • W Ritter et al.

    Orienting and habituation to auditory stimuli: A study of short-term changes in averaged evoked responses

    Electroencephalogr. Clin. Neurophysiol.

    (1968)
  • W.T Roth et al.

    Parameters of temporal recovery of the human auditory evoked potential

    Electroencephalogr. Clin. Neurophysiol.

    (1976)
  • H.H Rothman et al.

    Slow evoked cortical potentials and temporal features of stimulation

    Electroencephalogr. Clin. Neurophysiol.

    (1970)
  • H.G Vaughan et al.

    The sources of auditory evoked responses recorded from the human scalp

    Electroencephalogr. Clin. Neurophysiol.

    (1970)
  • W.R Webster

    The effects of repetitive stimulation on auditory evoked potentials

    Electroencephalogr. Clin. Neurophysiol.

    (1971)
  • J.R Wolpaw et al.

    The influence of stimulus intensity, contralateral masking and handedness on the temporal N1 and the T complex components of the auditory N1 wave, by John F. Connolly

    Electroencephalogr. Clin. Neurophysiol.

    (1994)
  • J.R Wolpaw et al.

    A temporal component of the auditory evoked response

    Electroencephalogr. Clin. Neurophysiol.

    (1975)
  • C.C Wood et al.

    Scalp distribution of human auditory evoked potentials. 11. Evidence for multiple sources and involvement of auditory cortex

    Electroencephalogr. Clin. Neurophysiol.

    (1982)
  • D.L Woods et al.

    Recovery cycles of event-related potentials in multiple detection tasks

    Electroencephalog. Clin. Neurophysiol.

    (1980)
  • D.L Woods et al.

    Generators of middle and long-latency auditory evoked potentials: Implications from studies of patients with bitemporal lesions

    Electroencephalogr. Clin. Neurophysiol.

    (1987)
  • D.L Woods et al.

    The recovery functions of the auditory event-related potentials during split-second discrimination

    Electroencephalogr. Clin. Neurophysiol.

    (1986)
  • D.L Woods et al.

    Bitemporal lesions dissociate auditory evoked potentials and perception

    Electroencephalogr. Clin. Neurophysiol.

    (1984)
  • S Zerlin et al.

    The variability of single evoked potentials in man

    Electroencephalogr. Clin. Neurophysiol.

    (1967)
  • M Alcaini et al.

    Two separate frontal components of the human auditory evoked response

    Psychophysiology

    (1994)
  • R.A Barry et al.

    Does the N100 evoked potential really habituate? Evidence from a paradigm appropriate to a clinical setting

    Int. J. Psychophysiol.

    (1992)
  • J.L Blom et al.

    An electrode cap tested

    Electroencephalogr. Clin. Neurophysiol.

    (1982)
  • T.W Bourbon et al.

    Habituation of auditory event-related potentials: a comparison of self initiated and automated stimulus trains

    Electroencephalogr. Clin. Neurophysiol.

    (1987)
  • T.W Budd et al.

    Facilitation of the N1 peak of the auditory ERP at short stimulus intervals

    NeuroReport

    (1994)
  • R.A Butler

    Effects of changes in stimulus frequency and intensity on habituation of the human vertex potential

    J. Acoust. Soc. Am.

    (1968)
  • Cited by (292)

    View all citing articles on Scopus
    View full text