Hoormann J 2000

Psychophysiology, 37 ~2000!, 2942. Cambridge University Press. Printed in the USA.
Copyright 2000 Society for Psychophysiological Research
Early attention effects in human

auditory-evoked potentials
JRG HOORMANN, MICHAEL FALKENSTEIN, and JOACHIM HOHNSBEIN

Institut fr Arbeitsphysiologie an der Universitt Dortmund ~IfADo!, Dortmund, Germany
Abstract
A fundamental question in attention theory concerns the earliest processing stages that can be modulated by selective
attention. A series of experiments is reported in which very early attention effects are found under specific conditions
in the frequency-following potential ~FFP!, a brain stem response to low-frequency tone stimuli. In two experiments,
stimuli of two different modalities were applied, and attention directed to one of the modalities. In two further
experiments, only auditory stimuli were presented. In the first of these last two experiments, a dichotic paradigm with
sustained attention to one ear was used, in the second a monotic paired-stimuli paradigm was used, in which the first
stimulus served as reference for the second one. Only in the last experiment significant attention effects were found in
the latency, but not in the amplitude of the FFP. The results show that a very early attention effect on the latency of the
FFP can be demonstrated, but only under highly specific conditions. The size and preconditions of the attention effect
suggest that it reflects subtle intramodal tuning mechanisms in the cochlea or in the lower brain stem.
Descriptors: Brain stem, Evoked potentials, Frequency-following potential, FFP, Attention
Peripheral Sensory Gating nerve ~Oatman, 1976; Oatman & Anderson, 1977! and the brain
stem ~Gabriel, Saltwick, & Miller, 1975; Olesen, Ashe, & Wein-
It has been long known from animal studies that efferent neurons berger, 1975!. In humans, evoked otoacoustic emissions ~EOAEs!
can modulate activity within afferent pathways, even at the level of have been used recently to demonstrate attention effects, because
the sensory receptor ~Granit, 1955; Hernandez-Peon, Scherrer, & these emissions are sensitive to mechanical alterations of the outer
Jouvet, 1956!. On the basis of these data, Hernandez-Peon ~1966! hair cells. The evidence of the studies on otoacoustic emissions is
proposed the peripheral filter model, which assumes that a de- not uniform. Earlier studies ~Avan & Bonfils, 1991; Freedman,
scending tonic inhibition can modulate sensory neurons. In fact, Adler, Waldo, Patchman, & Franks, 1990; Froehlich, Collet, Cha-
the hardware for such a filter exists, at least in the auditory system nal, & Morgon, 1990; Meric & Collet, 1992, 1993; Puel, Bonfils,
of mammals, namely the olivocochlear bundle ~OCB!. The OCB & Pujol, 1988! suffered from methodological problems, that is,
consists of two subsystems, a medial and a lateral subsystem comparison of passive with discrimination conditions. This criti-
~Guinan, 1988!. In the medial OCB, efferent fibers arise from cell cism does not apply to two more recent studies by Giard, Collet,
bodies in the superior olivary complex and terminate in the co- Bouchet, and Pernier ~1994! and Michie, LePage, Solowij, Haller,
chlea. This OCB appears to modulate the active micromechanical and Terry ~1996!. However, whereas Giard et al. ~1994! did find a
properties of the outer hair cells ~Mountain, 1980; Siegel & Kim, small ~average of 0.5 dB! but consistent effect of attention on
1982!. Scharf, Quigley, Aoki, Peachey, and Reeves ~1987!, who otoacoustic emissions in a narrow frequency band around the evok-
conducted psychophysical studies, postulated that this modulation ing frequency, Michie et al. ~1996! did not find such an effect.
is an active process to control specifically those frequencies that In sum, from animal studies there is ample evidence for pe-
are task relevant. Hence it is reasonable to assume that early at- ripheral gating even at the cochlear level, although attempts to
tention effects can be mediated by efferent pathways such as the demonstrate such gating for humans in recent psychoacoustical
olivocochlear bundle. Indeed, classical animal studies have dem- studies have been doubtful. To clarify this issue, since the early
onstrated effects of auditory attention at the level of the acoustical 1970s electroencephalographic data were used to study early at-
tention effects, which led to the results summarized in the following.
We thank Michael Smieja for his support with the hardware develop-
ment, Ludger Blanke for developing the software for EEG processing and Electrophysiology of Sensory Gating
presentation, Brigitta Voss for her helpful assistance with the data evalu-
ation, Christiane Westedt for producing the figures, and Carl R. Cavonius The physiological and psychological mechanisms of selective at-
for checking the manuscript. We also thank Steven Hackley for most valu- tention can be analyzed in humans by recordings of event-related
able comments and discussions during the preparation of this paper.
Address reprint requests to: J. Hoormann, Institut fr Arbeitsphysiolo-
potentials ~ERPs! from the scalp. Attention effects have been ob-
gie an der Universitt Dortmund, Ardeystr. 67, D-44139 Dortmund, Ger- served consistently for long-latency auditory ~and visual! ERPs.
many. E-mail: [email protected]. The most prominent attention-related phenomenon is a negative
29
30 J. Hoormann et al.
displacement for attended ~as opposed to unattended! auditory studies conducted by Oatman and co-workers ~Oatman, 1971, 1976;
stimuli, which has been called the N1 effect ~Hillyard, Hink, Oatman & Anderson, 1977, 1980; Glenn & Oatman, 1977! showed
Schwent, & Picton, 1973!, or the processing negativity ~Ntnen, clear attention effects on brain stem potentials in a bimodal para-
1982; Ntnen, Gaillard, & Mntysalo, 1978!. This phenomenon digm, when attention was directed to or away from the visual
is probably not a modulation of the N1, but an endogenous com- modality. Lukas ~1980, 1981! adapted the bimodal methodology
ponent that under certain conditions overlaps the N1. However, for experiments with humans. In both studies by Lukas, visual and
apart from the processing negativity, real amplitude modulations auditory stimuli occurred at random. In the first experiment of the
not only of the N1, but of other exogenous components ~e.g., the 1980 study, the subjects had either to count visual target letters or
P2! can be observed under conditions that are optimized for early to listen to tone pips of 1000 Hz. In the visual discrimination
selection ~e.g., Woldorff & Hillyard, 1991!. McCallum, Curry, condition, the amplitude of wave V was smaller, and its latency
Cooper, Pocock, and Papakostopoulos ~1983! even observed a longer, than in the listening condition. In a second experiment with
negative shift in the ERP to attended versus unattended stimuli as higher stimulus frequencies ~2000 and 8000 Hz!, similar attention
early as 15 ms after stimulus onset. In contrast, with a different effects were observed, but only for 2000 Hz.
task, Woldorff and Hillyard ~1991! found the ERP to be more The problem of this study is that tasks of different complexity
positive in the time range of 2050 ms for attended versus un- were assigned to the visual ~discrimination! and the auditory ~lis-
attended high-frequency tone pips ~P20-50!. By recording neuro- tening! modality. Hence in a second study ~Lukas, 1981! subjects
magnetic fields, Woldorff et al. ~1993! could locate the sources of had to detect targets in both modalities. The auditory stimuli were
the P20-50 as well as the later effect around 100 ms in the auditory 8000-Hz tone pips ~targets had a longer duration than nontargets!.
cortex. This result confirms the findings of Scherg and von Cra- No attention effect was found on the nontarget BER, but an effect
mon ~1985!, who located the sources of middle-latency compo- on wave I for the targets.
nents in the 2030-ms range in the primary and secondary auditory Lukas explained this targetnontarget difference by the differ-
cortex with the help of dipole estimation. ent frequency content of his targets and nontargets ~due to the
From these studies it is clear that selective attention can influ- different duration!, which hence stimulate slightly different areas
ence sensory processing in the auditory cortex. However, the ques- of the basilar membrane. However, this explanation is unlikely,
tion remains open whether any modulation of auditory processing because the power spectra of the two stimuli differ only slightly.
can operate at a more peripheral level, that is, in the brain stem or Furthermore, the specific delay of wave I ~and not of the later
even the cochlea. waves IIV! seems improbable, because wave I reflects the summed
activity of the eighth nerve action potentials ~Achor & Starr, 1980;
Moller, Jannetta, Bennett, & Moller, 1981!. A final critical point in
The Brain Stem Evoked Response (BER)
that study is the small number of averaged trials for targets.
Beginning in the early 1970s, considerable effort has been devoted Picton, Stapells, and Campbell ~1981! tried to replicate the
to investigate possible attention effects on very early auditory ERP results of Lukas ~1980! with an auditory listen0visual discrimina-
components in humans ~Picton & Hillyard, 1974; Picton, Hillyard, tion paradigm. They found no differences between the conditions
Galambos, & Schiff, 1971; Woods & Hillyard, 1978!. Picton et al. for waves I and V. In a preliminary report, Brix ~1984! measured
~1971! studied the cochlear nerve response, which has a latency of BERs to ~unspecified! clicks in 100 subjects who had either to
about 2 ms after click stimuli, in which high frequencies predom- attend to the clicks ~trying to count them! or to read a newspaper.
inate. These authors failed to find any attention-related modulation The author found a shorter interpeak latency between waves I and
of this very early auditory potential. Only a short time later the V in the attention condition. However, it remained unclear whether
far-field electrical activity from the human brain stem was discov- this shortening was due solely to a shortening of wave V latency.
ered by Jewett and Williston ~1971!. These brain stem evoked Also, the number of averaged epochs was rather small, and no
responses ~BERs!, which are evoked by short click stimuli, consist significance level for the effect was reported.
of five to seven positive components ~I to VII! with peak latencies Sommer ~1985! conducted a cross-modal attention study with
between 1.5 and 7 ms. Picton and Hillyard ~1974! examined BERs individuals with schizophrenia. For a subgroup of the patients who
during click intensity discrimination and during a passive control showed no electrodermal response to orienting stimuli ~nonrespond-
condition. They found no differences in any BER component be- ers!, the author found longer wave V latencies when visual rather
tween the attention ~discrimination! and control conditions, whereas than auditory stimuli were attended.
later ERP components showed a strong amplitude modulation. Connolly, Aubry, McGillivary, and Scott ~1989! used both intra-
Woods and Hillyard ~1978! considered the possibility that periph- and intermodal attention paradigms with high-frequency clicks as
eral gating might operate mainly in the context of ecologically auditory stimuli. These authors also found no significant change in
relevant, complex stimuli, such as speech. They presented clicks the BER waves across the attention conditions, which was again a
superimposed on different spoken prose passages presented to the failure to replicate Lukas ~1980! earlier results. However, their
two ears, while one of the passages had to be attended to. Again the Experiment 1 was not likely to have manipulated attention effec-
click-evoked BERs were unaffected by attention. Together these tively, because the visual task was too easy and coupled with the
results were interpreted as clear evidence against a peripheral gat- auditory task.
ing mechanism of auditory attention. Gregory, Heath, and Rosenberg ~1989! compared BERs in a
In the studies cited, attention was manipulated only within one passive condition versus an auditory versus a visual selective at-
~the auditory! modality. However, peripheral gating or filtering tention condition. In the two latter conditions, subjects had to
may be more effective if the difference between the attended and count occasional high-frequency targets ~8-kHz tones! among lower-
ignored stimuli is greater. Hence the optimum strategy to demon- frequency pips, or to count deviant-colored shapes among other
strate early attention effects could be to use intermodal paradigms shapes. The authors found no attention-related effects on the BER.
in which stimuli of two ~or more! modalities are presented, one of Hackley, Woldorff, and Hillyard ~1990! investigated cross-
which has to be attended. In fact, in the 1970s, a series of animal modal selective attention effects on retinal, myogenic, brain stem,
Early attention effects and evoked potentials 31
and cerebral potentials. The visual stimuli were flashes, the audi- main speech frequencies. Hence high-frequency click stimuli, such
tory stimuli high-frequency ~4000 Hz! tone pips. The target stimuli as also used by Woods and Hillyard, might not be optimally ap-
~about 5%! were less intense in both modalities. Whereas the longer- propriate. Consequently, brain stem responses that are sensitive to
latency components showed clear attention-related enhancements, lower frequencies might be more usefully examined.
neither the BER nor the early visual components ~electroretino-
gram, occipital components between 40 and 70 ms! were signifi-
The Frequency-Following Potential (FFP)
cantly affected by attention.
Bauer and Bayles ~1990! conducted an intermodal study in- The FFP ~Marsh, Smith, & Worden, 1972; Moushegian, Rupert, &
cluding the auditory and the somatosensory modalities. To focus Stillman, 1973! is a brain stem evoked response, which can be
attention within certain time periods, trains ~10s long! of clicks elicited by relatively low-frequency stimuli. The FFP is best re-
were presented to the left ear via headphones; the trains were corded from the vertex with reference at the mastoid contralateral
separated by pauses. subjects awaited either an auditory target ~a to the stimulated ear. However, because at least two sources or
500-Hz tone presented from an additional loudspeaker placed be- pathways for the FFP are assumed ~Marsh, Brown, & Smith, 1974!,
hind the subjects! or a somatosensory target presented with near- an ipsilateral reference is also frequently used ~e.g., Stillman, Crow,
threshold intensity at a random time during a train. Waves II and & Moushegian, 1978!. The FFP reproduces the frequency of the
V of the BER were significantly larger and peaked earlier in the waveform of the eliciting stimulus, and its onset is delayed by
auditory attention than in the somatosensory attention condition. about 6 ms relative to stimulus onset, which suggests a lower brain
The authors suggested two possible factors for their positive re- stem origin ~Galbraith & Brown, 1990! ~cf. also Figure 1!.
sults: ~i! the high task demands, that is, the use of low-intensity Marsh et al. ~1972! suggested the inferior colliculus as the
targets, which led to a highly focused attention, and ~ii! the large origin of the FFP. However, some years later, evidence was found
separation of targets and nontargets, that is, by frequency, time- that more peripheral brain stem sites, such as the superior olivary
course, and location. A further reason might be the good signal- complex or even the cochlear nucleus produce the FFP ~Gardi,
to-noise ratio, which is usually obtained when the subjects could Merzenich, & McKean, 1979; Hoormann, Falkenstein, & Hohns-
close their eyes, which is possible when visual stimuli are not bein, 1992a!. The FFP is thought to depend on phase-locking
involved. neural elements ~Gardi et al., 1979; Marsh et al., 1972; Stillman
Hirschhorn and Michie ~1990! attempted to compare inter- et al., 1978!, which code other stimulus properties than do neurons
modal with intramodal attention effects on the BER, using visual responsive to transients. The amplitude of the FFP is estimated by
and auditory stimuli. The auditory stimuli were loud clicks with calculating the fast Fourier transform and measuring the peak at
main frequencies around 1400 Hz ~nontargets! or 900 Hz ~targets; the stimulus frequency. The FFP latency is estimated by calculat-
4%!. The visual stimuli were nonsemantic shapes ~containing 25% ing the cross-correlation function between stimulus and FFP and
targets!. The stimuli were randomized. subjects had to attend to determining the latency of the maximum of this function. Hence a
one of the modalities either in a passive listen0look condition or in methodological advantage of the FFP over the BER is that ampli-
an active discrimination condition ~which was varied in difficulty!. tude and latency measures depend on the entire waveshape and not
Several comparisons were made for the amplitudes and latencies on a single peak, which enhances the accuracy of the measure-
of all BER waves. Among other significant findings, the latency of ment. The maximum amplitude FFP is obtained with tones around
wave I was prolonged during the difficult visual discrimination 350 Hz and is diminished strongly for frequencies beyond 500 Hz.
condition compared with the listen condition. This result replicated For very low frequencies ~100 to about 200 Hz! the FFP shows in
the early results of Lukas ~1980! and suggests an inhibition of all addition a strong second harmonic of the eliciting frequency ~Hoor-
auditory stimuli during highly focused visual attention. However, mann et al., 1992a!. The FFP can be also elicited by vowels and
this result was not reflected in the later BER waves. Also, the large natural speech stimuli ~Galbraith, Arbagey, Branski, Comerci, &
number of statistical tests ~80 altogether! greatly increases the Rector, 1995; Galbraith, Jhaveri, & Kuo, 1997!. Hence the FFP
possibility of reaching significance by chance. On the other hand, appears to be most sensitive to frequencies in the main speech
the only significant effects were found for waves I and II, which is range. If attention affects mainly speech information ~Woods &
unlikely to be due to chance. Hillyard, 1978!, the FFP should be the most appropriate tool for
In summary, whereas the intramodal BER studies so far showed assessing very early auditory attention effects.
no effects, the results of the intermodal BER studies are less con- Oatman and Anderson ~1977, 1980! used the FFP in their an-
sistent. Although most of the intramodal BER studies likewise imal studies. Cats were trained to perform a visual attention task.
failed to demonstrate any attention effects, a few intermodal stud- The visual stimuli consisted of two concentric rings presented with
ies did show effects on the BER waves I, II, and0or V. It cannot be varying stimulus onset asynchrony ~SOA!. The cats had to press a
excluded that the positive results were chance findings. On the lever after but not before the occurrence of the second ring. Tone
other hand, it is possible that under certain conditions positive bursts of different frequencies were presented simultaneously. The
effects may be real. It seems worthwhile to delimit such condi- authors demonstrated a clear suppression of the FFP amplitude for
tions. One basic precondition seems to be a sufficiently complex all stimulus frequencies during visual attention compared with pre-
task. Also, positive effects were observed mainly for latencies and posttest passive control conditions.
rather than for amplitudes of the BER waves. Finally, there seems In a first attempt to study attention effects on the FFP in hu-
to be a tendency for positive results when lower stimulus frequen- mans, Hillyard and Picton ~1979! used an intermodal paradigm in
cies are used. This tendency suggests the possibility that there are which attention was directed to the auditory or to the visual mo-
no peripheral gating mechanisms for high-frequency stimuli, but dality. Hillyard and Picton could find no effects on the amplitude
that there are for low-frequency stimuli such as speech ~which are of the FFP despite significant effects on later components. Possible
ecologically more valid for humans!. This possibility has already effects on FFP latency were not mentioned.
been mentioned by Woods and Hillyard ~1978!. Unfortunately, the Galbraith and Kane ~1993! also investigated intermodal atten-
frequencies that contribute to the BER are usually higher than the tion effects on the FFP. In a visual attention condition, subjects
Figure 1. Example for a pure tone burst ~320 Hz; upper panels! that elicits a frequency-following potential ~FFP; lower panels,
averaged response of a single S!. The left panels present the time course, the right panels the frequency contents of both stimulus and
response. The response is delayed by about 6 ms relative to the stimulus and reproduces its frequency.
performed a complex visual task, while ignoring simultaneously overlapping activity of two FFP generators, which had been claimed
presented 230-Hz tone pips. In an auditory attention condition with by Stillman et al. ~1978!.
closed eyes, the subjects had to silently count complex auditory In a further study, Galbraith and Doan ~1995! presented stimuli
stimuli interspersed among the tones. Although the cortical-evoked with a different quality, namely tones and missing fundamental
potential to the tones showed a clear attention-related modulation, ~MF! stimuli. MF stimuli have the same pitch as the pure tones of
no attention effect on the FFP ~averaged across 750 sweeps! was the fundamental frequency, but do not physically contain the fun-
found. The authors attributed their negative results to the very high damental frequency. MF stimuli have been shown to evoke similar
discriminability of targets from nontargets in the auditory attention FFPs as tone stimuli ~Galbraith, 1994; Greenberg, 1980!. Galbraith
condition, which may have rendered a strong focusing of attention and Doan ~1995! assumed that different anatomical and physio-
unnecessary. Moreover, both of these studies used low stimulus logical properties are involved in the coding of MF and pure tones
rates ~which might be ineffective in terms of maximum focussing and could hence yield different FFP results with respect to atten-
of attention! because the investigators intended to record long- tion. For both types of stimuli, the targets differed from nontargets
latency ERPs in addition to FFPs. Finally, the number of sweeps either by a lower intensity ~8-dB difference! or by a longer dura-
~about 750! for the FFP in the latter study may be too low for tion ~75 vs. 25 ms!. Attention was directed to one of the ears in a
detecting small differences. block; the direction of attention was changed between blocks. The
In a further study, Galbraith and Arroyo ~1993! used the FFP in FFP amplitude determination was based on the entire length of the
an intramodal selective attention task. They presented different waveshape ~i.e., not divided into two parts as in Galbraith & Ar-
tone bursts ~with frequencies of 200 and 400 Hz, respectively! royo, 1993!. Again a complex result was found, namely that the
randomly and asynchronously to the two ears, the interstimulus FFP was only larger for attended ~vs. unattended! MF stimuli,
interval ~ISI! was randomized between 120 and 320 ms. In differ- whereas the FFP was even smaller for attended tones, particularly
ent blocks, subjects had to attend to one particular ear and fre- when intensity was the discriminative cue. The latter discrimina-
quency ~e.g., high tones in the left ear! and press a key to each tion was also the most difficult one, as shown in the d9 scores. This
target, which was less intense than the other tones. Because the pattern of results is hard to explain and may be partly due to
FFP waveshape appeared to consist of two parts, the FFP ampli- differences in task difficulty between the conditions, as was also
tude was determined for the first and second parts separately. A discussed by the authors. However, the fact that the task with the
complex result was found: for attended 400-Hz stimuli, the am- most difficult discrimination yielded a negative attention effect is
plitude of the second part of the FFP was reduced compared with puzzling. If the FFP effects for the MF and the easy duration
the first part, whereas the opposite was found for nonattended discrimination were due to attention, they should become even
400-Hz stimuli. For 200-Hz stimuli, the result for the ignored larger for a harder discrimination, because a harder discrimination
stimuli was in the opposite direction. The authors stated cautiously requires even more attention. In addition, the overall discrimina-
that some form of attention-related modulation may be occurring bility of targets and nontargets proved to be high, as revealed in a
in the lower brain stem. The complex result was explained by the mean hit rate of 90%. Also, no effects on late ERP components in
the N1 range were reported. Hence it is doubtful whether attention acoustic signal by about 30 ms relative to the electromagnetic
was effectively manipulated in that study, and a nonattention ex- signal, hence excluding stimulus artifacts in the response window.
planation for the results should be considered. A further problem The device consists of a mu-metal shielded low-frequency speaker,
may have been that only 500 sweeps were averaged to obtain the which was coupled via an exponential horn to a long PVC tube.
FFP in each condition, a number that does not seem to be sufficient The tube was led into a sound-attenuated, electrically shielded
considering the small size of the expected attention effects. chamber. The proximal end of the tube was coupled to a circu-
In sum, the results of attention effects on FFPs are contradic- maural ear cushion. The total distance between the speaker mem-
tory, and no latency results were reported. However, the two most brane and the eardrum was 9.2 m, which causes the 30-ms delay of
recent reports of the Galbraith group yielded at least partially the auditory stimulus. The electrical signal to the speaker was
positive results. prefiltered with an active filter to equalize the frequency charac-
teristic of the entire system. Distortions of the signal at the tube-
exit induced by the entire system were negligible, which was
The Present Study
monitored via a miniature microphone in the ear cushion. Here the
The present study was aimed to continue the research on attention second harmonic of the signal was less than 55 dB ~rel. funda-
effects with the FFP, because early attention effects are expected to mental! in the range of 1281088 Hz. ~For further technical details
occur in the low frequencies in the speech band. Besides looking cf. Hoormann et al., 1992b.!
for amplitude effects, we were particularly interested in latency Second, several techniques were used to enhance the signal-to-
effects, because these effects have not been reported up to now, but noise ratio: ~i! to reduce the total electromyogram ~EMG!, the
can be expected, because latency effects were reported several subjects lay on a vacuum mattress shaped exactly to their body,
times for the click-evoked BER. Moreover, the variance of FFP which enabled them to relax very comfortably; ~ii! special low-
latency is smaller than that of FFP amplitude ~Hoormann et al., noise preamplifiers were used and the gain was adjusted individ-
1992a!. We studied FFPs with several paradigms; the bimodal ually; ~iii! the passband was filtered very sharply with 10-pole
paradigm was thought to be the most promising, because the to- Butterworth filters; and ~iv! a sufficient number of artifact-free
be-attended channels ~the modalities! have maximum separation sweeps, that is, at least 2,500 per condition, were averaged to
~cf. Connolly et al., 1989!. To further enhance the focusing of obtain the FFP. To resolve very small latency differences, the
attention, the discriminability of targets and nontargets was made electroencephalogram ~EEG! was sampled with 250 kHz per
difficult ~Woldorff & Hillyard, 1991!. In two experiments we used channel.
a bimodal paradigm with one modality attended. Attention was
either sustained for longer time intervals ~Experiment 1! or the
EXPERIMENT 1
relevant modality was cued in each trial ~Experiment 2!. To study
possibly different distraction mechanisms for different modalities, This was a cross-modal attention study with sustained attention to
the modality pairs were varied ~auditory-visual or auditory- one modality.
somatosensory! in Experiment 2. In two further experiments, only
auditory stimuli were used ~unimodal paradigm!. In the first of
Methods
these ~Experiment 3!, a dichotic listening paradigm with sustained
attention was used; during one block subjects had to attend exclu- Subjects
sively to one ear and press a key after the occurrence of an occa- Twenty healthy subjects ~10 men, 10 women, mean age 5 25
sional deviant. As in the study of Galbraith and Doan ~1995!, pure years; naive as to the purpose of the experiment! participated. All
tones and missing fundamental stimuli were used. To control for subjects had normal ~or corrected-to-normal! vision. Before the
the effectiveness of the attention manipulation, middle- and long- EEG session their hearing threshold was measured at 10 test fre-
latency ERPs ~up to 200 ms! were also measured. In the second of quencies by an audiometer and found to be within normal limits
these experiments ~Experiment 4! ~short report in Hoormann, Falk- ~65 dB; American National Standards Institute, 1970!. The sub-
enstein, & Hohnsbein, 1994! monotic paired-stimuli paradigm with jects were paid 15 DM per hour for participating.
short intervals between the pairs was used, in which the first stim-
ulus served as reference for the second, which could be the same Stimuli
or ~occasionally! slightly different. subjects had to focus attention The auditory stimuli were low-frequency tone bursts presented to
very strongly after the first stimulus to be able to distinguish for the left ear via the auditory delay line. The stimulus frequency was
the subsequent stimulus a target from a nontarget and could then 320 Hz ~nontargets! or 337 Hz ~targets!. The stimulus envelope
reduce the focussing of attention until the next trial, since the first was symmetrically Gaussian-shaped enclosing a plateau ~for de-
stimulus was always the same. This paradigm was thought to lead tails cf. Hoormann, Voss, Falkenstein, & Hohnsbein, 1997!; the
to a strong phasic focusing of attention to the critical time segment. total stimulus length was 10 cycles ~i.e., circa 30-ms duration as
were all stimulus durations in all four experiments!, including two
cycles rise and two cycles fall time. The pitch difference between
Methods
targets and nontargets was chosen in pilot studies to establish
General Remarks roughly a 75% probability of correct target identification. The
Because we expected only small ~if any! attention effects, careful intensity of the stimuli was adjusted individually to 80 dB HL.
method refinements were carried out in advance to enhance the The high intensity was chosen to obtain maximum amplitudes of
quality of the FFP. First, we avoided the use of earphones because the FFPs ~Hoormann et al., 1992a!. To avoid the overlap of FFPs
they either lead to stimulus artifacts in the response or, if shielded from both lateral brain stem sites ~as elicited by binaural stimula-
by mu-metal, they may distort the signal ~Batra, Kuwada, & Maher, tion!, the right ear was masked with white noise of 65 dB SPL. The
1986; Hoormann, Falkenstein, & Hohnsbein, 1992b!. Instead, for visual stimuli were LED flashes of 50 ms length. The LED was a
each stimulated ear we used a separate tube device that delayed the duo-LED ~capable of emitting green or red light or a mixture of
both!, which was controlled such that either pure green flashes ~non- Results and Discussion
targets! or green flashes with a slight addition of red ~targets! were
A total of 81% of the auditory targets and 78% of the visual targets
emitted. The intensity was held constant for nontargets and targets.
were identified correctly. Although the difference between these
The hue difference between targets and nontargets was chosen in
levels was not significant, the difference between correct auditory
pilot studies to establish a roughly 75% correct target identification.
~415 ms! and visual ~462 ms! reaction times ~RTs! was significant,
F~1,19! 5 23.84, p , .0001. The mean FFP amplitude was 380 nV.
Procedure
The FFP was larger with contralateral ~410 nV! than with ipsilat-
Each session had three parts: audiometry, attention part, and pas-
eral reference ~351 nV!, F~1,19! 5 20.35, p , .0005. No further
sive part. In the attention part, auditory and visual stimuli were
amplitude effects were found, F , 1. The mean FFP latency was
presented in random sequence with an ISI randomized around
6.34 ms. No significant effects at all were found for the FFP
320 ms ~240 400 ms!. The stimuli were presented in blocks of
latency. An explanation for the absence of any attention effect on
100 stimuli; before each block a visual or an auditory cue was
the FFP may be the high intensity of the stimuli, because the
given that indicated the to-be-attended modality during the follow-
olivocochlear bundle exerts efferent control mainly for low-intensity
ing block. The visual cue was a pure red stimulus from the duo-
stimuli ~e.g., Siegel and Kim, 1982!. Hence softer stimuli were
LED, the auditory cue was a 1-kHz tone. Visual and auditory cues
used in the subsequent experiments. A further reason may be that
were given in alternation, that is, the subjects had to change the
attention cannot be focused strongly enough for an entire block of
attended modality every block. About 320 ms after the cue the
32 s ~in our case! to influence early auditory processing. Hence in
stimulus train began. In the attended modality, 5% of the nontar-
a subsequent experiment the focusing of attention was directed to
gets were replaced randomly by targets, whereas in the non-
short time periods only. Furthermore, the somatosensory modality
attended modality only nontargets were presented to avoid the
was included as an alternative distractor modality, in addition to
possibility of distraction of attention by targets in the nonattended
the visual modality.
modality. The task was to press a key after each target. The first
three sweeps were excluded from averaging in each block. After
each 10 min a short break was inserted. Altogether the attention EXPERIMENT 2
part had a length of about 80 min. In the passive part, only auditory
stimuli were presented, with the same ISI as in the attention part. In this experiment stimulus pairs were presented to the same or to
The subjects had simply to relax, and ignore the stimuli. Also in different modalities. The S1-S2-interval ~SSI! was short, whereas
the passive part 3,000 artifact-free sweeps were collected. the interpair0intertrial interval ~ITI! was long. The first stimulus
specified the to-be-attended modality: when the second stimulus
EEG Recording and Data Evaluation was presented to the same modality as the first stimulus, it was
The EEG was recorded between Cz and right ~contralateral! as attended, otherwise it was unattended. Hence, subjects had to focus
well as left ~ipsilateral! mastoid; the forehead served as ground. attention during the SSI, whereas they could relax during the ITI.
Electrode resistance was kept below 2 kV. The subjects were asked
to relax and to avoid any tension in the head muscles. The pre- Methods
amplifier gain was adjusted individually between 392,000 and
758,000 to make best use of the full range of the preamplifiers. The Subjects
EEG was analog-filtered by high-pass ~100 Hz! and low-pass Twelve healthy naive subjects ~6 men, 6 women, mean age 5 24
~600 Hz! 10-pole Butterworth filters, and sampled with 250 kHz years! participated. All subjects had normal ~or corrected-to-
per channel ~i.e., sampling interval of 4 ms; further processing of normal! vision and hearing thresholds ~see Experiment 1!. The
the data was conducted with the same resolution!. A total of 3,000 subjects were paid 15 DM per hour for participating.
artifact-free sweeps were collected and averaged for the nontargets
in the two ~contra- and ipsilateral! channels; artifacts ~trials that Stimuli and Procedure
exceeded the preamplifier range! were discarded online. The re- Several short sessions were administered to induce a high perfor-
cording window had a length of 55 ms beginning 8 ms before mance level and to obtain a sufficient number of sweeps per con-
~acoustic! stimulus onset. The amplitude of the averaged FFP was dition. All subjects participated in six sessions, each on a separate
estimated by its spectral power, which was calculated using the day. In three of the sessions, auditory and visual stimuli were
DADiSP software package. The latency of the FFP was estimated presented, in the three remaining sessions auditory and somato-
in the time domain by the latency of the maximum of the cross- sensory stimuli were presented. The order of sessions was balanced
correlation function of the response with the acoustic stimulus across subjects. The auditory stimuli were presented monaurally
~using BMDP2T; for further details see Hoormann et al., 1992a!. via the delay line, the visual stimuli via a duo-LED, as described
Both were fed through the same set of analog filters, hence group in Experiment 1. The nontargets and targets had the same charac-
delay and phase shift were the same for both. An artificial ear teristics as in Experiment 1. The somatosensory stimuli were ap-
~Brel & Kjaer 4153, Naerum, Denmark! and an impulse sound plied to the skin of the left lower leg ~about 20 cm above the outer
level meter ~Brel & Kjaer 2209! were used to measure the sound ankle! by a mechanical stimulation device, which avoided electro-
pressure level and the reference signal for the cross-correlation magnetic artifacts in the chamber. The stimulation device con-
with the stimulus. Within a multivariate design, the dependent sisted of a small plastic frame fastened to the leg in which two
variables amplitude and latency and the independent within-factors movable pins are mounted at a distance of 4 cm. Both pins could
recording channel ~contralateral, ipsilateral! and attention ~audito- be moved ~about 2-mm extension! independently by two 5-m long
ry, visual, passive! were analyzed by a two-way multivariate analy- cables in a tube and were controlled from outside the chamber.
sis of variance ~MANOVA!. The degrees of freedom for the Two possible stimuli were applied by this device: a nontarget was
independent factor attention with three levels were adjusted after a short ~500 ms long! touch of one pin, a target was a synchronous
Geisser and Greenhouse ~1958!. touch of both pins. The pitch difference ~auditory stimuli!, the hue
difference ~visual stimuli!, and the touch difference ~somatosen- Altogether the results of the cross-modal Experiments 1 and 2
sory stimuli! between targets and nontargets were determined in revealed no significant effects of focusing of attention on the am-
pilot studies to establish a roughly 75% probability of correct plitude or latency of the FFP. In the subsequent experiments we
target identification. The intensity of the auditory stimuli was set to investigated intramodal attention effects with auditory stimuli only.
65 dB HL, which was much less than in Experiment 1. As in
Experiment 1, targets occurred with 5% probability in the attended
EXPERIMENT 3
modality. The stimuli were presented as pairs ~S1 and S2! with an
ISI of 600 ms between S1 and S2, and an ITI of 2,400 ms between In this first intramodal experiment, we applied an equivalent of the
the pairs. The first stimulus in a pair specified the to-be-attended classical Hillyard paradigm ~Hillyard et al., 1973!, that is, auditory
modality, the second stimulus was presented to the to-be-attended stimuli were presented at a fast rate to the two ears with different
or to the to-be-unattended modality with equal probability. As in frequencies; one ear had to be attended to. A crucial point in the
Experiment 1, the subjects had to press a key after a target. Hillyard paradigm is a large separation of the attended and non-
FFPs were recorded to all auditory stimuli, which could be cue, attended channels, and a difficult targetnontarget discrimination.
attended auditory stimulus ~with a preceding auditory cue!, or Hence we used both ear position and frequency as channels, and
unattended auditory stimulus ~with a preceding visual or somato- made the discriminability of targets from nontargets difficult. Be-
sensory cue!. As in Experiment 1, both contra- and ipsilateral cause Galbraith and Doan ~1995! found positive results mainly
references were used. In each condition 3,000 artifact-free sweeps with missing fundamental stimuli, and we also intended to repli-
were averaged and evaluated as described above. The maximum of cate parts of that study, we did not only apply pure tones as stimuli,
the cross-correlation function was evaluated in addition as a mea- but also missing fundamental stimuli. Finally, to control the effec-
sure for the quality of the averaged FFP. Long-latency auditory as tiveness of the attention procedure, middle- and long-latency
well as visual and somatosensory evoked potentials ~EPs! were auditory-evoked potentials ~AEPs! were recorded in addition to
computed in addition. The late auditory EPs were not further eval- the FFP.
uated, because the factor attention is confounded with the ISI,
which is known to influence the amplitude of the N1 ~e.g., Davis,
Methods
Mast, Yoshie, & Zerlin, 1966!: for attended tones ~auditory cue
preceding auditory stimulus! the time interval between successive Subjects
auditory stimuli is short ~600 ms!, which is likely to suppress N1 Twelve healthy female subjects with a mean age of 24 years par-
amplitude, whereas for nonattended tones ~nonauditory cue pre- ticipated in this experiment. One subject had to be excluded, be-
ceding auditory stimulus! it is long ~more than 3,000 ms!, which is cause she did not meet the hearing threshold criteria ~see Experiment
likely to enhance N1 amplitude. Hence the ISI effect on N1 am- 1!; one further subject had no measurable FFP. The remaining 10
plitude would counteract and obscure a possible attention effect. subjects had a mean age of 23 years.
The dependent variables FFP amplitude and latency were analyzed
by a three-way repeated-measures MANOVA with the independent Stimuli and Procedure
within-factors recording channel ~contralateral, ipsilateral!, con- All subjects participated in three sessions, each on a separate day.
current modality ~visual, somatosensory!, and condition ~cue, au- On the first day, the audiometry was conducted, and the subjects
ditory attention, auditory nonattention!. practiced the main task. Particularly they were trained to direct
attention exclusively to one ear and switch attention to the other
ear after a cue. On the next 2 days, pure tone stimuli and missing
Results and Discussion
fundamental stimuli were presented. The order of the tone and
The maximum value of the cross-correlation function was higher missing fundamental days was balanced across subjects. The au-
for the contralateral ~0.80! than for the ipsilateral recordings ~0.73!, ditory stimuli ~circa 30 ms duration! were presented to both ears
which shows the better quality of the contralateral recordings. Also via two independent auditory delay lines ~cf. Experiments 1 and 2!.
the FFP appeared to be larger with contralateral ~275 nV! than For the pure tone stimulation, 240-Hz tones ~nontargets! or 200-Hz
with ipsilateral reference ~239 nV!; however, this difference was tones ~targets! were presented to the right ear, and 380-Hz tones
not significant. No significant result was obtained for attention, ~nontargets! or 340-Hz tones ~targets! to the left ear. These fre-
though there was an FFP latency difference in the expected direc- quencies were chosen so as to yield maximum channel separation
tion ~attended 6.54 ms, nonattended 6.58 ms!. ~maximum frequency difference; Hillyard et al., 1973! within the
In sum, there was no significant attention effect on any FFP most sensitive frequency range of the FFP ~Hoormann et al., 1992a!.
parameter. Hence, even if attention is strongly focused to a par- The nontargets were presented in random sequence to the ears; in
ticular modality over a split-second period, there are no significant the attended ear, about 8% of the nontargets were replaced ran-
attention effects on FFP latency or amplitude. The possibility that domly by targets. In this experiment no targets were presented to
the different ISIs between subsequent attended and nonattended the nonattended ear to avoid ~automatic! distraction of attention to
stimuli ~see above! have influenced the FFP result is very unlikely, the nonattended ear. Indeed, the optimal focusing of attention seemed
because the FFP is invariant to different ISIs down to 100 ms ~e.g., so important to us that we accepted the asymmetry of the exper-
Glaser, Suter, Dasheiff, & Goldberg, 1976!. A further reason may imental design with respect to the distribution of targets between
be the switching of attention across modalities, which has to be the two ears. The ISI was randomized between 200 and 400 ms.
expected from trial to trial, and which has to be done every other The intensity of the 380-Hz tone was 65 dB SL; the intensity of the
trial on average. After 600 ms, ISI attention can be expected to be 240-Hz tone was adjusted individually until it was perceived with
focused on the to-be-attended modality; however, a cost of switch- the same loudness as the 380-Hz tone. The stimulus trains were
ing between modalities cannot be excluded. This may have led to divided into blocks of 400 stimuli. Each block was started by the
a relative dilution of attention, which may have cancelled the subjects with a key press, which was followed by a cue ~a 1-kHz,
expected effect due to focusing of attention to a narrow time segment. 50-dB SPL tone of 2-s length! that was given to the ear that had to
be attended in the following block. The order of right0left cues0 Results and Discussion
blocks was regular: Four right cues0blocks were followed by four
left cues0blocks; this sequence was repeated 300 ms after the cue The RTs for pure tones were 459 ms ~right ear stimulation! and
the stimulus train began. Again the subjects had to press a key after 420 ms ~left ear!, and for missing fundamental stimuli 407 ms ~right
each target in the attended ear. The first three sweeps of each block ear! and 391 ms ~left ear!. The RT differences between right and left
were excluded from averaging. were significant, F~1,9! 5 31.65, p , .0005, as were the differ-
For the missing fundamental stimulation, complex stimuli were ences between the RTs after tones and missing fundamental stimuli,
produced by superpositioning third, fourth, and fifth harmonics of F~1,9! 5 18.58, p , .005; the interaction was not significant. A to-
fundamentals, which correspond to the pure tones used in the pure tal of 8.4% ~right ear! and 5.0% ~left ear! of the tone targets were
tone stimulation. Hence the right ear was stimulated with a com- missed, and 6.1% ~right ear! and 7.7% ~left ear! of the missing
plex of 720, 960, and 1200 Hz ~nontargets! or 600, 800, and fundamental targets. The main effects were not significant, F , 1,
1000 Hz ~targets!, the left ear with a complex of 1140, 1520, and although the interaction was significant, F~1,9! 5 5.74, p , .05!.
1900 Hz ~nontargets! or 1020, 1360, and 1700 Hz ~targets!. In all The FFP was clearly larger in the vertical than in the horizontal
other respects the stimulation procedure was the same as for the channel for the 240-Hz stimuli; no other significant amplitude or
pure tones. latency effects were found. For the vertical FFP ~including both
Because we found no attention specific differences for FFPs stimulus frequencies! no significant amplitude or latency effect
after ipsi- versus contralateral stimulation in Experiments 1 and 2, was found. Figure 2 shows the FFPs after attended and non-
we used a vertical and horizontal derivation as proposed by attended stimuli for two subjects.
Galbraith ~1994! for better distinction of the different sources of The N100 following tone stimuli was larger at Cz than at the
tone and missing fundamental FFPs. The vertical derivation was two other electrodes, F~2,18! 5 6.29, p , .05, E 5 0.72, which is
Cz versus left mastoid and the horizontal derivation was right the usual finding. A similar, though weaker, effect was found after
versus left mastoid. The FFPs were recorded to all nontarget stim- missing fundamental stimuli. More important, the N100 after tone
uli, which could be attended or unattended, and evaluated as de- stimuli was larger for attended ~21.7 mV! than for unattended
scribed above; 2,500 artifact-free sweeps were averaged separately stimuli ~21.0 mV!, F~1,9! 5 16.41, p , .005 ~Figure 3!. A similar
for each condition. effect was found for missing fundamental stimuli: the N100 was
In addition to the FFPs, middle-latency and late-evoked poten- larger for attended ~22.4 mV! than for unattended stimuli
tials were recorded from the three midline electrodes, Fz, Cz, and ~21.3 mV!, F~1,9! 5 27.83, p , .001. As also seen in Figure 3, not
Pz, with the left mastoid as reference. The middle-latency poten- only the N100, but also the subsequent ERP segment appears to be
tials were low-pass filtered with 250 Hz and sampled with 800 Hz, shifted in the negative direction for attended versus nonattended
the late potentials were low-pass filtered with 60 Hz and sampled stimuli.
with 200 Hz. The group delays introduced by the low pass filters For the middle-latency potentials ~Figure 4! no significant la-
were calculated and subtracted from the raw latency values. Again, tency effects were found. For the amplitudes the P11 and P32
2,500 sweeps were collected per condition; electrooculogram ~EOG! showed no significant attention effects, though the P32 was nu-
and slow artifacts were eliminated offline, which resulted in about merically less positive after attended stimuli, as also seen in Fig-
2,100 artifact-free sweeps per condition. The averages were ure 4. The N22 was more negative after attended than after
smoothed digitally ~Ruchkin & Glaser, 1978! below 117 Hz ~for nonattended stimuli, F~1,9! 5 5.64, p , .05. As also seen in
middle-latency potentials! and below 17.6 Hz ~for late potentials!, Figure 4, not only the N22, but a broader time window between 15
and then corrected for EOG artifacts ~Verleger, Gasser, & Mcks, and 50 ms ~averaged across stimuli! appears to be shifted in the
1982!. The middle-latency potentials revealed three deflections in negative direction for attended compared with nonattended stimuli.
the latency range between 10 and 50 ms after stimulus onset, called However, this effect failed to reach significance, p 5 .14.
P11, N22, and P32. In the late range, the N100 was detected as the These results demonstrate the effectiveness of the attention
largest relative minimum in the window 80160 ms after stimulus manipulation, as revealed in the amplitude effects on N100 and
onset. All deflections were measured relative to electrical zero N22. The latter effect shows a negative displacement of the N22
~i.e., zero voltage at the differential amplifiers!; for the middle- after attended versus unattended stimuli, which is similar to the
latency potentials the mean amplitude in the window 1550 ms findings of McCallum et al. ~1983!. On the other hand, the present
poststimulus was computed in addition ~see below!. data are different from the results of Woldorff et al. ~1987!, be-
The FFP could not be measured reliably in the horizontal chan- cause we could not detect any positive shift such as the P20-50.
nel for the 380-Hz missing fundamental stimulus. Hence, it was mea- This result may be due to task differences, that is, much lower
sured in the vertical channel for both frequencies and its amplitude stimulus frequencies and larger ISIs in the present task versus that
and latency ~dependent variables! analysed ~separately for pure tones of Woldorff et al. . In any case, the present results show coherent
and missing fundamental stimuli! by a two-way repeated-measures effects, namely negative displacements for attended versus un-
MANOVA with the dependent within-factors frequency0ear ~2400 attended stimuli in middle and late AEPs.
right, 3800left!, and attention ~attended, nonattended!. In addition, In contrast, no effect at all was seen for the FFP, that is, we could
the 240-Hz data were evaluated with the design channel ~vertical, not find larger FFP amplitudes after attended than after unattended
horizontal! and attention ~attended, nonattended!. For the middle- tone or MF stimuli, a finding that contradicts the results of Gal-
latency potentials only the data for the tones were evaluated ~be- braith and Doan ~1995!. The negative result is not likely due to the
cause the missing fundamental data showed no clear peaks! with the fact that discriminability was lower in our experiment compared with
design electrode ~Fz, Cz, Pz!, frequency, and attention. The ampli- Galbraith and Doans; attention should be focused even more ef-
tude and latency of the N100 were evaluated separately for tones fectively when the discriminability of targets and nontargets is lower.
and missing fundamental stimuli with the design electrode ~Fz, Cz, Also, we could show a strong modulation of the N100 with atten-
Pz!, frequency, and attention. Where appropriate the degrees of tion, which proves the effectiveness of the attention manipulation.
freedom were adjusted after Geisser and Greenhouse ~1958!. However, the discrepancy may be due to the use of a frequency dif-
Figure 2. Examples of two typical individual frequency-following potentials ~FFPs! after attended and nonattended stimuli ~upper part
subject EZ, lower part subject BV!. Left panels: FFPs after 380 Hz stimulus, right panels: FFPs after 240 Hz stimulus. att 5 attended
ear; non-att 5 nonattended ear; S 5 stimulus onset.
ference in the present study instead of duration0intensity differ- cussing of attention. In our next experiment, a unimodal monaural
ences in the study of Galbraith and Doan ~1995!, because early paradigm was administered, in which attention was highly focused
attention mechanisms may depend on the nature of the task required over very short intervals.
~frequency vs. intensity vs. duration discrimination!. Such possible
mechanisms are discussed below ~Experiment 4!.
EXPERIMENT 4
Hence, it appears that a classical unimodal condition with sus-
tained attention to one ear, which does produce the well-known To focus attention over short time intervals, a double stimulus
attention effects on the N100, does not affect very early process- paradigm was applied. In contrast to the cross-modal study with
ing, as reflected in the FFP. This finding may be due to the fact that stimulus pairs ~Experiment 2!, the first stimulus was not a cue, but
a sustained allocation of attention allows only for suboptimal fo- served as the reference for the second stimulus. The interval be-
Figure 3. Grand means ~n 5 10 subjects; Experiment 3! for the late auditory-evoked potentials ~up to 200 ms after stimulus onset!
after 240-Hz tones ~upper panel! and after 380-Hz tones ~lower panel! for attended ~ATT; heavy lines! and unattended stimuli ~NATT;
thin lines!. The later part of the event-related potential ~ERP! is more negative after attended than after nonattended stimuli.
tween first and second stimulus was chosen to be very short, which presented with a short time interval under passive conditions, a
should permit a very effective focusing of attention restricted to suppression of the amplitude of a middle-latency auditory poten-
this period ~Ntnen, 1990!. When auditory stimulus pairs are tial, the P50, is usually found for the second compared with the
Figure 4. Grand means ~n 5 10 subjects; Experiment 3! for the middle-latency auditory-evoked potentials ~up to 70 ms after stimulus
onset! after 240-Hz tones ~upper panel! and after 380-Hz tones ~lower panel! for attended ~heavy lines! and unattended stimuli ~thin
lines!. The event-related potential ~ERP! is more negative after attended than after nonattended stimuli in the time segment 1550 ms
~averaged across stimuli!.
first stimulus ~Freedman, Adler, Waldo, Patchman, & Franks, Results and Discussion
1983; Waldo & Freedman, 1986!. Guterman, Josiassen, and
Bashore ~1992! could show that this P50 suppression was not No significant effects at all were found for the FFP amplitude.
evident when subjects attended to the second stimulus to detect ~The amplitude of the FFP was numerically larger for the contra-
a target. Guterman et al. concluded that this cancellation of the lateral than for ipsilateral channel, however, the difference did not
P50 suppression reflects an active attention control process, which reach the significance level.! The maxima of the cross-correlation
already influences rather early auditory processing. Hence, we functions between stimuli and responses ~that were calculated to
expected to have a chance to find even earlier attention effects determine the latencies of the FFPs! were between 0.79 ~contra-
with this paradigm. lateral! and 0.74 ~ipsilateral!. Of course, these values are highly
significant and demonstrate the high quality of the latency estima-
tions. No significant main effects and interactions of the factor
Methods recording channel were found for the FFP latency. The results
for the other two factors on FFP latency are summarized in Fig-
Subjects ure 5. The factor condition revealed a strong tendency, F~1,10! 5
Twelve healthy young subjects ~5 men, 7 women! with normal 4.71, p 5 .055, for a shorter FFP latency in the attention than in the
auditory thresholds participated. One female subject had to be ignore condition. The factor stimulus was highly significant,
excluded because no FFP could be measured. The remaining 11 F~1,10! 5 17.68, p , .005. Most importantly, the condition 3
subjects had a mean age of 21 years ~range, 1829 years!. stimulus interaction was also significant, F~1,10! 5 5.34, p , .05.
The latter result allowed the computation of simple effects for both
Stimuli and Procedure attention conditions. This analysis revealed that the FFP latency
The stimuli were low-frequency tone bursts ~c. 30 ms duration, 85 was highly significantly shorter for the second stimulus ~6.32 6
dB SPL! presented to the left ear via the auditory delay line ~see 0.061 ms @mean 6 SEM #! than for the first stimulus ~6.37 6
above!; the right ear was masked with white noise of 65 dB SPL. 0.063 ms! in the attention condition, F~1,10! 5 13.54, p , .005,
The duration and envelope of the stimuli was as described for the whereas there was no latency difference at all between the two
earlier experiments. The stimuli were presented in pairs with a stimuli in the ignore condition ~S1: 6.37 6 0.063 ms, S2: 6.37 6
very short S1-S2 interval ~300 ms!, and an interpair interval of 0.066 ms!. The latency reduction is quantitatively small. However,
about 1,850 ms ~randomized between 1,300 and 2,400 ms!. The the reliability of this effect is not only revealed by the clear AN-
first stimulus of each pair had a frequency of 340 Hz and served as OVA results, but also is underlined by the fact that in 22 compar-
the reference for the second stimulus, which could be either iden- isons ~11 subjects, 2 sides! only a single latency showed a slight
tical with the first stimulus ~nontarget!, or a 358-Hz stimulus ~tar- increase.
get!. Targets occurred on between 0.5% and 5% of the presentations The fact that in the ignore condition no latency difference was
in different blocks ~see below!. The pitch differences between seen between the FFPs to the first and the second stimulus shows
targets and nontargets were chosen in pilot studies to establish a
roughly 75% probability of correct target identification. ~The crit-
ical pitch difference was relatively small because of the immediate
succession of the to-be-discriminated stimuli.! The task of the
subjects was to count the targets covertly and report the result at
the end of a block, which consisted of 150 stimulus pairs. After
each block, feedback was given as to the quality of the perfor-
mance by giving the actual number of targets in the past block. The
average block length was 5 min. These blocks were repeated until
a sufficient number of artifact-free sweeps were collected for each
of the 340-Hz stimuli ~see below!; for this task about 30 blocks
were necessary on average. Because the presentation of pairs may
have an automatic suppressive effect on the FFP to the second
stimulus ~as found for the P50!, we established a control condition,
in which the same stimulus pairs were presented, while the sub-
jects had to ignore the stimuli and relax. In the control condition no
targets were presented. Again, blocks of 150 pairs were given until
the same number of artifact-free sweeps was collected as in the
attention condition.
FFPs were recorded to all 340-Hz stimuli, that is, reference
stimuli and nontargets in both attention and ignore conditions. The
FFP was measured at Cz relative to the contralateral and the ipsi-
lateral mastoid. A total of 3,000 artifact-free sweeps were averaged
per condition. The FFP amplitude and latency were evaluated as
described above.
The amplitude and latency ~dependent variables! of the FFP Figure 5. Mean latencies ~n 5 11 subjects; Experiment 4! for the frequency-
was analyzed by a three-way repeated-measures MANOVA with following potentials ~FFPs! after the first and the second stimulus of the
the independent within-factors recording channel ~contralateral, paired stimuli in the attention and the ignore condition. The FFP latency is
ipsilateral!, stimulus ~first, second!, and condition ~attend, significantly shorter after the second stimulus in the attention, but not in the
ignore!. ignore condition.
that automatic facilitation cannot be the reason for the reduced ~Experiment 3!, although the task was difficult and the attention
latency in the attention condition. Also the reduced arousal in the manipulation effective, as seen in the N100 and N22 effects. Hence,
ignore condition ~indeed, some of the subjects were actually asleep it may be concluded that mechanisms supporting sustained atten-
in this condition! is unlikely to be the reason for the latency re- tion to one ear0frequency are likewise not influenced by peripheral
duction in the attention condition, because the FFP latency was gating, as reflected in the FFP. Only when attention was directed
exactly the same for the first stimulus in both the attention and the exclusively to one ear, and focused to short time segments to cope
ignore condition. Any effect due to a difference in arousal should with a difficult frequency discrimination task, could early attention
have influenced the responses after the first stimulus as well as effects be demonstrated ~Experiment 4!.
after the second stimulus, which was not the case. It seems also We therefore conclude that early auditory attention effects are
unlikely that general arousal has changed between the first and the evident mainly in unimodal situations with unilateral stimuli, when
second stimulus ~i.e., within 300 ms in the attention condition! for attention is highly focused to a restricted time interval. Hence
several thousand times in a row. Hence, these results suggest that peripheral gating does not support a general facilitation or inhibi-
in the unimodal monaural condition the processing of the second tion of an entire stimulus modality, but serves subtle intramodal
auditory stimuli in pairs that were highly attended, were facilitated mechanisms, which may be necessary for a better localization
on the level of the lower brain stem or even at a more peripheral and0or frequency discrimination of task-relevant stimuli. The pre-
level. cise origin of the observed FFP latency effect within the auditory
The effect is, however, numerically very small, and clearly pathway is still unclear, and can be located, at present, only be-
irrelevant for reaction tasks. It may be speculated what kind of tween the cochlea and the FFP source. Because we provided ev-
mechanism could be affected by this small peripheral effect. One idence for the cochlear nucleus as the source of the FFP ~Hoormann
possibility is that the effect supports the auditory localization of et al., 1992a!, the origin of the effect is probably the cochlea or the
low-frequency stimuli in the horizontal plane ~lateralization!, be- cochlear nucleus. Recent studies of otoacoustic emissions suggest
cause latency differences in the range of 50 ms lead to detectable control of outer hair cells via the olivocochlear bundle ~Giard
lateralization changes. Moreover, Hall ~1979! assumed that FFPs et al., 1994; Meric & Collet, 1992; Michie et al., 1996!. In partic-
reflect neural information relevant for lateralization. Hence it may ular, it has been demonstrated that activating the medial olivoco-
be possible that our FFP latency result reflects the action of a chlear system by contralateral acoustic stimulation contributes to
neural mechanism that underlies the sharpening of spatial atten- shortening otoacoustic emission latency ~Giraud, Perrin, Chry-
tion. A further possibility is that the effect reflects a sharpening of Croze, Chays, & Collet, 1996; Giraud et al., 1997!. This result may
pitch discrimination, because highly accurate pitch discrimination be relevant to the FFP-latency shortening we observed. Hence, we
is in fact required to cope with the task. This idea is supported by favor the view that the attention effect on FFPs has its origin in the
the fact that the auditory efferent fibers, particularly the olivoco- cochlea.
chlear bundle, support a sharpening of the frequency tuning in the An issue that has to be discussed in more detail is whether
cochlea by modulating the active micromechanical properties of the double-stimulus paradigms ~Experiments 2 and 4! stimulate
the outer hair cells ~Mountain, 1980; Scharf et al., 1987; Siegel & an involuntary capture of attention ~exogenous orienting! rather
Kim, 1982!. than a voluntary allocation of attention ~endogenous orienting!
~Spence & Driver, 1994! and whether the latency shortening of
the FFP in Experiment 4 is due to exogenous or endog-
GENERAL DISCUSSION AND CONCLUSION
enous orienting of attention or, alternatively, to general arousal
In none of the four experiments reported we could find any effect rather than to attention. For visual stimuli, Rafal, Henik, and
of attention on FFP amplitude. This finding is in line with the Smith ~1991! found that subcortical structures are only involved
negative findings of Hillyard and Picton ~1979! and Galbraith and in exogenous attention. Schrger and Eimer ~1996! showed
Kane ~1993!, but disagrees with some positive results of Galbraith late negative modulations of the ERP ~Nd! after peripheral
and Arroyo ~1993!1 and Galbraith and Doan ~1995!. In a quasi- stimuli presented at the same position as an immediately preced-
replication of the experiment by Galbraith and Doan ~1995!, we ing cue stimulus even when the cue was uninformative or mis-
could demonstrate the effectiveness of our attention manipulation leading ~invalid!. However, the Nd was seen only in an active
by effects on late ~N100! and middle-latency AEPs ~N22!, whereas condition, but not in an ignore condition. This observation sug-
no attention effect on the FFP was found. Hence we conclude that gests that exogenous attention is ~i! not fully automatic ~as
with cross-modal and with intramodal paradigms there is no effect also claimed by Spence & Driver, 1994!, and ~ii! occurs in
on FFP amplitude, even when attention is focused strongly to paradigms with paired auditory stimuli ~cue and imperative stim-
narrow time segments. ulus!, as applied in our study. Taken together, the results of the
As to FFP latency, there were no significant attention effects in two cited studies support our view that the early attention ef-
cross-modal sustained and cued attention paradigms ~Experiments fects we found in Experiment 4 may be mainly due to exog-
1 and 2!. Hence, it may be concluded tentatively that peripheral enous attention. However, because of the difficulty of the task
gating as reflected in the FFP is not relevant or possible for inter- in our paradigm, a substantial contribution of endogenous at-
modal attention mechanisms, which facilitate or attenuate an entire tention, that is, a voluntary sustained attention to the left ear and
modality. additional allocation of attention after S1, seems possible. Con-
As to the intramodal experiments ~Experiments 3 and 4!, no cerning the possible contribution of arousal, we cannot defi-
early attention effect was seen for a sustained attention paradigm nitely exclude some influence on the observed latency effect.
However, two findings seem to contradict this interpretation: ~i!
1
Only in the attention condition of Experiment 4 was the FFP-
Contrary to these authors we could not find any differential attention
effect for the first vs. second half of the FFP. The reason for this discrep-
latency after the second stimulus shorter in comparison with all
ancy is not clear to us, but it is a fact that the signal-to-noise ratio of the other conditions. The latency after the first stimulus was virtu-
FFPs was much higher in our experiment. ally the same in the control and the attention condition. If gen-
eral arousal indeed played a role, these results meant, as discussed also. Such an effect, that is, a difference of the FFP latency
above, that, on the one hand, arousal was the same in both between the attended and nonattended condition, was absent.2
conditions ~sleep and attention!, but that, on the other hand, To us this absence suggests that the latency reduction of our
arousal changed within the SOA of 300 ms for 3,000 times in a Experiment 4 was not an arousal effect.
row. ~ii! Possibly even more relevant may be the result of our To shed more light on the contribution of both varieties of
Experiment 2, which was a cross-modal task with stimulus pairs. attention and of arousal on the early latency effect, we plan a
Indeed, in a similar cross-modal task, Low, Larson, Burke, and further FFP experiment with pure endogenous attention and pure
Hackley ~1996! demonstrated an immediate arousal effect ~alert- exogenous attention as proposed by Schrger and Eimer ~1996!.
ing! for the R50 component ~50 ms after the evoking stimulus!
of the eyeblink reflex, when the evoking visual stimulus was
preceded by an accessory tone by only 40 ms. Therefore, one 2
Even in the condition with a pair of two auditory stimuli in our
should expect an immediate arousal effect in our Experiment 2 Experiment 2 there was no immediate arousal effect.
REFERENCES
Achor, L. J., & Starr, A. ~1980!. Auditory brain stem responses in the cat. Geisser, S., & Greenhouse, S. W. ~1958!. An extension of Boxs results on
I. Intracranial and extracranial recordings. Electroencephalography and the use of the F distribution in multivariate analysis. Annals of Math-
Clinical Neurophysiology, 48, 154173. ematics and Statistics, 29, 885891.
American National Standards Institute ~1970!. Specifications for Audiom- Giard, M. H., Collet, L., Bouchet, P., & Pernier, J. ~1994!. Auditory selec-
eters. 53.6-1969. New York: ANSI. tive attention in the human cochlea. Brain Research, 633, 353356.
Avan, P., & Bonfils, P. ~1991!. Analysis of possible interactions of an Giraud, A. -L., Perrin, E., Chry-Croze, S., Chays, A., & Collet, L. ~1996!.
attentional task with cochlear micromechanics. Hearing Research, 57, Contralateral acoustic stimulation induces a phase advance in evoked
269275. otoacoustic emissions in human. Hearing Research, 94, 54 62.
Batra, R., Kuwada, S., & Maher, V. L. ~1986!. The frequency-following Giraud, A. -L., Wable, J., Chays, A., Collet, L., & Chry-Croze, S. ~1997!.
response to continuous tones in humans. Hearing Research, 21, 167177. Influence of contralateral noise on distortion product latency in hu-
Bauer, L. O., & Bayles, R. L. ~1990!. Precortical filtering and selective mans: Is the medial olivocochlear efferent system involved? Journal of
attention: an evoked potential analysis. Biological Psychology, 30, 2133. the Acoustical Society of America, 102, 22192227.
Brix, R. ~1984!. The influence of attention on the auditory brain stem Glaser, E. M., Suter, C. M., Dasheiff, R., & Goldberg, A. ~1976!. The
evoked response. Acta Otolaryngologica (Stockh.), 98, 8992. human frequency-following response: Is behavior during continuous
Connolly, J. F., Aubry, K., McGillivary, N., & Scott, D. W. ~1989!. Human tone and tone burst stimulation. Electroencephalography and Clinical
brainstem auditory evoked potentials fail to provide evidence of effer- Neurophysiology, 40, 2532.
ent modulation of auditory input during attentional tasks. Psychophys- Glenn, J. F., & Oatman, L. C. ~1977!. Effects of visual attention on the
iology, 26, 292303. latency of auditory evoked potentials. Experimental Neurology, 57,
Davis, H., Mast, T., Yoshie, N., & Zerlin, S. ~1966!. The slow response of 34 40.
the human cortex to auditory stimuli: Recovery process. Electroenceph- Granit, R. ~1955!. Receptors and sensory perception. New Haven. CT: Yale
alography and Clinical Neurophysiology, 21, 105113. University Press.
Freedman, R., Adler, L. E., Waldo, M. C., Patchman, E., & Franks, R. D. Greenberg, S. ~1980!. Temporal neural coding of pitch and vowel quality.
~1983!. Neurophysiological evidence for a defect in inhibitory path- UCL A Working Papers in Phonetics, 52, 1183.
ways in schizophrenia: Comparison of medicated and drug free pa- Gregory, S. D., Heath, J. A., & Rosenberg, M. E. ~1989!. Does selective
tients. Biological Psychiatry, 18, 537551. attention influence the brain-stem auditory evoked potential? Electro-
Froehlich, P., Collet, L., Chanal, J. M., & Morgon, A. ~1990!. Variability of encephalography and Clinical Neurophysiology, 73, 557560.
the influence of a visual task on the active micromechanical properties Guinan, J. J. ~1988!. Physiology of the olivocochlear efferents. In J. Syka
of the cochlea. Brain Research, 508, 286288. ~Ed.!, Auditory pathway structure and function ~pp. 253267!. New
Gabriel, M., Saltwick, S. E., & Miller, J. D. ~1975!. Conditioning and York: Plenum Press.
reversal of short-latency multiple-unit responses in the rabbit medial Guterman, Y., Josiassen, R., & Bashore, T. R. ~1992!. Attentional influence
geniculate nucleus. Science, 189, 11081109. on the P50 component of the auditory event-related potential. Inter-
Galbraith, G. C. ~1994!. Two-channel brain-stem frequency-following re- national Journal of Psychophysiology, 12, 197209.
sponses to pure tone and missing fundamental stimuli. Electroenceph- Hackley, S. A., Woldorff, M., & Hillyard, S. A. ~1990!. Cross-modal se-
alography and Clinical Neurophysiology, 92, 321330. lective attention effects on retinal, myogenic, brainstem, and cerebral
Galbraith, G. C., Arbagey, P. W., Branski, R., Comerci, N., & Rector, P. M. evoked potentials. Psychophysiology, 27, 195208.
~1995!. Intelligible speech encoded in the human frequency-following Hall, J. W. ~1979!. Auditory brainstem frequency-following responses to
response. NeuroReport, 6, 23632367. waveform envelope periodicity. Science, 205, 12971299.
Galbraith, G. C., & Arroyo, C. ~1993!. Selective attention and brainstem Hernandez-Peon, R. ~1966!. Physiological mechanisms in attention. In R. W.
frequency-following responses. Biological Psychology, 37, 322. Russell ~Ed.!, Frontiers in physiological psychology ~pp. 121144!.
Galbraith, G. C., & Brown, W. S. ~1990!. Cross-correlation and latency New York: Academic Press.
compensation analysis of click-evoked and frequency-following re- Hernandez-Peon, R., Scherrer, H., & Jouvet, M. ~1956!. Modifications of
sponses. Electroencephalography and Clinical Neurophysiology, 77, electrical activity in the cochlear nucleus during attention in unanes-
295308. thetized cats. Science, 123, 331332.
Galbraith, G. C., & Doan, B. Q. ~1995!. Brainstem frequency-following Hillyard, S. A., Hink, R. F., Schwent, V. L., & Picton, T. W. ~1973!.
and behavioral responses during selective attention to pure tone and Electrical signs of selective attention in the human brain. Science, 182,
missing fundamental stimuli. International Journal of Psychophysiol- 177180.
ogy, 19, 203214. Hillyard, S. A., & Picton, T. W. ~1979!. Event-related brain potentials and
Galbraith, G. C., Jhaveri, S. P., & Kuo, J. ~1997!. Speech-evoked brainstem selective information processing in man. In J. E. Desmedt ~Ed.!, Cog-
frequency-following responses during verbal transformations due to nitive components in cerebral event-related potentials and selective
word repetition. Electroencephalography and Clinical Neurophysiol- attention. Progress in clinical neurophysiology ~Vol. 6, pp. 152!. Ba-
ogy, 102, 4653. sel, Switzerland: Karger.
Galbraith, G. C., & Kane, J. M. ~1993!. Brainstem frequency-following Hirschhorn, T. N., & Michie, P. T. ~1990!. Brainstem auditory evoked
responses and cortical event-related potentials during attention. Per- potentials ~BAEPs! and selective attention revisited. Psychophysiology,
ceptual and Motor Skills, 76, 12311241. 27, 495512.
Gardi, J., Merzenich, M., & McKean, C. ~1979!. Origins of the scalp-recorded Hoormann, J., Falkenstein, M., & Hohnsbein, J. ~1992a!. The human
frequency-following response in the cat. Audiology, 18, 353381. frequency-following response ~FFR!: Sources of variability and rela-
tion to the click-evoked brainstem response ~BER!. Hearing Research, tioning in the paralyzed cat. Journal of Neurophysiology, 38, 1114
59, 179188. 1139.
Hoormann, J., Falkenstein, M., & Hohnsbein, J. ~1992b!. An improved Picton, T. W., & Hillyard, S. A. ~1974!. Human auditory evoked potentials.
stimulation technique for the human frequency-following response. Amer- II: Effects of attention. Electroencephalography and Clinical Neuro-
ican Journal of EEG Technology, 32, 213224. physiology, 36, 191200.
Hoormann, J., Falkenstein, M., & Hohnsbein, J. ~1994!. Effect of selective Picton, T. W., Hillyard, S. A., Galambos, R., & Schiff, M. ~1971!. Human
attention on the latency of human frequency-following potentials. Neu- auditory attention: A central or peripheral process? Science, 171, 351353.
roReport, 5, 16091612. Picton, T. W., Stapells, D. R., & Campbell, K. B. ~1981!. Auditory evoked
Hoormann, J., Voss, B., Falkenstein, M., & Hohnsbein, J. ~1997!. A method potentials from the human cochlea and brainstem. Journal of Oto-
to improve the latency estimation of the frequency-following potential laryngology, 9~Suppl.!, 1 41.
~FFP!. Electroencephalography and Clinical Neurophysiology, 103, Puel, J. L., Bonfils, P., & Pujol, R. ~1988!. Selective attention modifies the
326329. active micromechanical properties of the cochlea. Brain Research, 447,
Jewett, D. L., & Williston, J. S. ~1971!. Auditory-evoked far fields averaged 380383.
from the scalp of humans. Brain, 94, 681-696. Rafal, R. D., Henik, A., & Smith, J. ~1991!. Extrageniculate contributions
Low, K. A., Larson, S. L., Burke, J., & Hackley, S. A. ~1996!. Alerting to reflex visual orienting in normal humans: A temporal hemifield
effects on choice reaction time and the photic eyeblink reflex. Electro- advantage. Journal of Cognitive Neuroscience, 3, 322328.
encephalography and Clinical Neurophysiology, 98, 385393. Ruchkin, D. S., & Glaser, E. M. ~1978!. Some simple digital filters for
Lukas, J. H. ~1980!. Human auditory attention: The olivocochlear bundle examination of CNV and P300 waveforms on a single trial basis. In
may function as a peripheral filter. Psychophysiology, 17, 444 452. D. A. Otto ~Ed.), Multidisciplinary perspectives in event-related po-
Lukas, J. H. ~1981!. The role of efferent inhibition in human auditory tential research ~pp. 579581!. Washington, DC: U.S. Government
attention: An examination of the auditory brainstem potentials. Inter- Printing Office.
national Journal of Neuroscience, 12, 137145. Scharf, B., Quigley, S., Aoki, C., Peachey, N., & Reeves, A. ~1987!. Fo-
Marsh, J. T., Brown, W. S., & Smith, J. C. ~1974!. Differential brainstem cussed auditory attention and frequency selectivity. Perception and
pathways for the conduction of auditory frequency-following responses. Psychophysics, 42, 215223.
Electroencephalography and Clinical Neurophysiology, 36, 415 424. Scherg, M., & von Cramon, D. ~1985!. Two bilateral sources of the late
Marsh, J. T., Smith, J. C., & Worden, F. G. ~1972!. Receptor and neural AEP as identified by a spatio-temporal dipole model. Electroencepha-
responses in auditory masking of low frequency tones. Electroenceph- lography and Clinical Neurophysiology, 62, 32 44.
alography and Clinical Neurophysiology, 32, 6374. Schrger, E., & Eimer, M. ~1996!. Effects of lateralized cues on the pro-
McCallum, W. C., Curry, S. H., Cooper, R., Pocock, P. V., & Papakosto- cessing of lateralized auditory stimuli. Biological Psychology, 43,
poulos, D. ~1983!. Brain event-related potentials as indicators of early 203226.
selective processes in auditory target localization. Psychophysiology, Siegel, J. H., & Kim, D. O. ~1982!. Efferent neural control of cochlear
20, 117. mechanics? Olivo-cochlear bundle stimulation affects cochlear biome-
Meric, C., & Collet, L. ~1992!. Visual attention and evoked otoacoustic chanical nonlinearity. Hearing Research, 6, 171182.
emissions: A slight but real effect. International Journal of Psycho- Sommer, W. ~1985!. Selective attention differentially affects brainstem
physiology, 12, 233235. auditory evoked potentials of electrodermal responders and nonrespond-
Meric, C., & Collet, L. ~1993!. Comparative influence of repeated mea- ers. Psychiatry Research, 16, 227232.
surement and of attention on evoked otoacoustic emissions. Acta Oto- Spence, C. J., & Driver, J. ~1994!. Covert spatial orienting in audition:
laryngologica (Stockh.), 113, 471 477. Exogenous and endogenous mechanisms. Journal of Experimental Psy-
Michie, P. T., LePage, E., Solowij, N., Haller, M., & Terry, L. ~1996!. chology: Human Perception and Performance, 20, 555574.
Evoked otoacoustic emissions and auditory selective attention. Hearing Stillman, R. D., Crow, G., & Moushegian, G. ~1978!. Components of the
Research, 98, 54 67. frequency-following potential in man. Electroencephalography and Clin-
Moller, A. R., Jannetta, P. J., Bennett, M., & Moller, M. B. ~1981!. Intra- ical Neurophysiology, 44, 438 446.
cranially recorded responses from human auditory nerve: New insights Verleger, R., Gasser, T., & Mcks, J. ~1982!. Correction of EOG artifacts in
into the origin of brain stem evoked potentials ~BSEPs!. Electroenceph- event-related potentials of the EEG: Aspects of reliability and validity.
alography and Clinical Neurophysiology, 52, 1827. Psychophysiology, 19, 472 480.
Mountain, D. ~1980!. Changes in endolymphatic potential and crossed Waldo, M. C., & Freedman, R. ~1986!. Gating of auditory evoked re-
olivocochlear bundle stimulation alter cochlear mechanics. Science, sponses in normal college students. Psychiatry Research, 19, 233239.
210, 7172. Woldorff, M. G., Gallen, C. C., Hampson, S. A., Hillyard, S. A., Pantev,
Moushegian, G., Rupert, A. L., & Stillman, R. D. ~1973!. Scalp-recorded C., Sobel, D., & Bloom, F. E. ~1993!. Modulation of early sensory
early responses in man to frequencies in the speech range. Electro- processing in human auditory cortex during auditory selective atten-
encephalography and Clinical Neurophysiology, 35, 665 667. tion. Proceedings of the National Academy of Sciences (USA), 90,
Ntnen, R. ~1982!. Processing negativity: An evoked-potential reflection 87228726.
of selective attention. Psychological Bulletin, 92, 605 640. Woldorff, M. G., Hansen, J. C., & Hillyard, S. A. ~1987!. Evidence for
Ntnen, R. ~1990!. The role of attention in auditory information process- effects of selective attention in the mid-latency range of the human
ing as revealed by event-related potentials and other brain measures of auditory event-related potential. In R. Johnson, Jr., J. W. Rohrbaugh, R.
cognitive function. Behavioral and Brain Sciences, 13, 201288. Parasuraman ~Eds.!, Current trends in event-related potential research
Ntnen, R., Gaillard, A. W. K., & Mntysalo, S. ~1978!. Early selective (EEG Suppl. 40) ~pp. 146154!. Amsterdam, Netherlands: Elsevier
attention effect on evoked potential reinterpreted. Acta Psychologica, Science Publishers B.V.
42, 313329. Woldorff, M. G., & Hillyard, S. A. ~1991!. Modulation of early auditory
Oatman, L. C. ~1971!. Role of visual attention on auditory evoked poten- processing during selective listening to rapidly presented tones. Elec-
tials in unanesthetized cats. Experimental Neurology, 32, 341356. troencephalography and Clinical Neurophysiology, 79, 170191.
Oatman, L. C. ~1976!. Effects of visual attention on the intensity of audi- Woods, D. L., & Hillyard, S. A. ~1978!. Attention at the cocktail party:
tory evoked potentials. Experimental Neurology, 51, 4153. Brainstem evoked responses reveal no peripheral gating. In D. A. Otto
Oatman, L. C., & Anderson, B. W. ~1977!. Effects of visual attention on tone ~Ed.!, Multidisciplinary perspectives in event-related brain potential
burst evoked auditory potentials. Experimental Neurology, 57, 200211. research ~pp. 230233!. Washington, DC: U.S. Government Publishing
Oatman, L. C., & Anderson, B. W. ~1980!. Suppression of the auditory Office.
frequency-following response during visual attention. Electroencepha-
lography and Clinical Neurophysiology, 49, 314322.
Olesen, T. D., Ashe, J. H., & Weinberger, N. M. ~1975!. Modification of
auditory and somatosensory system activity during pupillary condi- ~Received July 23, 1998; Accepted March 9, 1999!

Hoormann J 2000

Uploaded by

Copyright:

Available Formats

Hoormann J 2000

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hoormann J 2000

Uploaded by

Copyright:

Available Formats

Psychophysiology, 37 ~2000!, 2942. Cambridge University Press. Printed in the USA.

Copyright 2000 Society for Psychophysiological Research

Early attention effects in human

JRG HOORMANN, MICHAEL FALKENSTEIN, and JOACHIM HOHNSBEIN

You might also like