NeuroImage 16, 1052–1061 (2002)doi:10.1006/nimg.2002.1142

Tryptophan Depletion Effects on EEG and MEG Responses SuggestSerotonergic Modulation of Auditory Involuntary Attention in Humans

Jyrki Ahveninen,*,†,‡,†† Seppo Kahkonen,*,†,†† Sirpa Pennanen,§ Jyrki Liesivuori,¶Risto J. Ilmoniemi,†,†† and Iiro P. Jaaskelainen� ,**

*Cognitive Brain Research Unit and **Apperception & Cortical Dynamics, Department of Psychology, P.O. Box 13, University of Helsinki,Helsinki, FIN-00014, Finland; ‡Department of Neurology, University of Helsinki, Helsinki, Finland; †BioMag Laboratory, Engineering

Centre, Helsinki University Central Hospital, Helsinki, Finland; §Occupational Hygiene and Toxicology Section, Kuopio RegionalInstitute of Occupational Health, Kuopio, Finland; ¶Department of Pharmacology and Toxicology, University of Kuopio,

Kuopio, Finland; �Massachusetts General Hospital-NMR Center, Harvard Medical School, Charlestown,Massachusetts 02129; and ††Helsinki Brain Research Center, Helsinki, Finland

Involuntary attention shifting, i.e., detecting andorienting to unexpected stimulus changes, may be al-tered at low brain serotonin (5-hydroxytryptamine;5-HT) levels. This was studied in 13 healthy subjects(21–30 years old; 6 females) by using a dietary chal-lenge, acute tryptophan depletion (ATD), which de-creases 5-HT synthesis in the brain. Five hours afteringestion of either ATD or control mixture (random-ized, double-blinded, crossover design), brain responsesindexing involuntary attention were measured withsimultaneous 64-channel electroencephalography (EEG)and 122-channel magnetoencephalography (MEG). Dur-ing the measurement, the subjects were instructed todiscriminate equiprobable 200- and 400-ms tones bypressing one of two buttons rapidly. Occasionally, thefrequency of the tones changed (10% increase/de-crease), causing involuntary attention shifting. ATDsignificantly lowered plasma tryptophan concentra-tions (total tryptophan decreased by 75%, free trypto-phan decreased by 35%). As compared to the controlcondition, ATD reduced the amplitude of the deviant-tone N2 wave, including the overlapping mismatchnegativity (MMN) and N2b subcomponents, which aresuggested to reflect change detection in the brain. TheEEG results were accompanied by a significant in-crease in the peak latency of the magnetic counterpartof MMN. However, no ATD effects were observed in P3to task-irrelevant frequency change. Reaction time(RT) to deviants per se was not significantly affected,but RT in trials succeeding the deviant-frequencytones was increased by ATD, which suggested im-paired reorienting to the task-relevant activity. Inconclusion, the results suggest that decreased level ofcentral 5-HT function after ATD may decrease invol-untary attention shifting to task-irrelevant soundchanges and thus modulate resource allocation to thetask-relevant activity. © 2002 Elsevier Science (USA)

10521053-8119/02 $35.00© 2002 Elsevier Science (USA)All rights reserved.

Key Words: attention; auditory cortex, event-relatedpotentials; electroencephalography; magnetoencepha-lography; mismatch negativity; P3; P3a; RON; seroto-nin; tryptophan depletion.

INTRODUCTION

Involuntary attention shifting enables rapid orientingto unexpected changes in the environment. Deficits inthis fundamental function could result in unresponsive-ness to potentially harmful events, whereas disinhibitionof attention shifting may distract goal-directed behavior(Ahveninen et al., 2000). The neurochemical regulationof involuntary attention, enabling us to selectively sus-tain or shift the focus of attention to task-relevant aspectsof information, is not yet fully known; however, mono-aminergic systems innervating the forebrain have beenthought to play a central role (Robbins et al., 1989).

A substantial amount of evidence associating themonoamine neurotransmitter serotonin (5-hydroxy-tryptamine; 5-HT) to attention has been obtained fromanimal experiments and human studies. Changes inprefrontal 5-HT and dopamine utilization have beenassociated with a poor ability to sustain attention inrats (Puumala and Sirvio, 1998). These results havebeen supported by indirect evidence arising from hu-man studies on different clinical subject groups. Atten-tion deficits might be common in abusers of methyl-enedioxymethamphetamine, i.e., “ecstasy” (McCann etal., 1999), which has been shown to result in specificand irreversible deficits in 5-HT neurons in rats(Ricaurte et al., 1992). At the same time, an inability tosustain attention is often observed in depression (Liottiand Mayberg, 2001), and, notably, the prefrontal ab-normalities thought to underlie such attention defi-cits (Dolan et al., 1994) can possibly be reversed byusing serotonin uptake inhibitors such as paroxetine

Received Oc

er 26, 2001 tob

(Kennedy et al., 2001). There is also evidence thatserotonin uptake inhibitors may improve attentionalperformance in healthy human subjects (Nathan et al.,2000). However, the relationship between the level of5-HT turnover and attention might not be linear. Inchildren with attention-deficit disorder, the impair-ments in sustained attention might correlate with anexcess amount of peripheral 5-HT metabolites (Oades,2000); moreover, a recent result suggests that experi-mentally reduced central 5-HT levels might, in healthysubjects, improve the ability to focus attention(Schmitt et al., 2000).

The neural basis of involuntary attention can bestudied by combining behavioral and neurophysiologi-cal measurements. In a recently developed paradigm(Schroger and Wolff, 1998a,b), subjects are instructedto discriminate between two sound durations. Occa-sionally, the frequency of the tones changes. Increasesin reaction time (RT) and decreases in hit rate (HR)caused by such task-irrelevant changes provide a be-havioral index of involuntary attention shifting,whereas event-related potentials (ERP) and magneticfields, stimulus-locked and averaged epochs of electro-encephalogram (EEG) and magnetoencephalogram(MEG), reflect the underlying cerebral events. TheERP component mismatch negativity (MMN) elicitedat 100–300 ms after the task-irrelevant frequencychange, and its MEG counterpart (MMNm), might re-flect “automatic” change detection required for invol-untary attention shifting, whereas the overlappingN2b component may indicate “active” detection of stim-ulus changes (Naatanen et al., 1982; Naatanen, 1992).These overlapping components, MMN/MMNm andN2b, are in turn often referred to as the N2 deflec-tion. N2 (MMN/N2b) is followed by the positive P3acomponent, possibly reflecting a later stage of noveltyprocessing, associated with the evaluation of the stim-ulus change for subsequent behavioral action (Fried-man et al., 2001). Finally, a slow negativity termedreorienting negativity (RON), purported to reflect orient-ing back to the task-relevant activity, is elicited at 400–600 ms after stimulus onset (Schroger and Wolff, 1998a).

The neurochemical bases of brain response associ-ated with involuntary attention are not fully known.Previous results, however, suggest that MMN/MMNmgeneration may be decreased by different glutamate-receptor antagonists in monkeys (Javitt et al., 1996)and by the cholinergic antagonist scopolamine in hu-mans (Pekkonen et al., 2001). Temazepine, an agonistof �-aminobutyric acid subtype a (GABAA) receptors,might suppress MMN (Hirvonen et al., 1998), and theGABAA-receptor agonist lorazepam has been shown toreduce the N2b amplitude (Berchou et al., 1986). Theinhibitory effects of acute ethanol, based on decreasedglutamatergic exitability and increased GABAA inhibi-tion (for a review, see Harris, 1999), in turn, diminishbehavioral distractibility, reduce the MMN and P3a

amplitude, and prolong the MMN peak latency(Jaaskelainen et al., 1999; for a review, see Jaaskel-ainen et al., 1996). There is some evidence also of N2bpeak latency increase after acute ethanol administra-tion (Teo and Ferguson, 1986). Finally, the dopamineD2-receptor antagonist haloperidol has been found toincrease MMN elicited by task-irrelevant deviants dur-ing a dichotic listening task, presumably reflecting im-paired focused attention to the task-relevant channel(Kahkonen et al., 2001). There is also evidence thathaloperidol might suppress P3a and RON (Kahkonenet al., 2002). No evidence of noradrenergic modulationof MMN has been found so far (Duncan and Kaye,1987; Mervaala et al., 1993).

Despite the presumed role of 5-HT in attention def-icits, few studies on serotonergic modulation of invol-untary attention have been published. One of the mainlines of research on auditory ERP and the 5-HT systemhas, instead, concentrated on intensity dependence ofthe N1/P2 components, suggesting that low 5-HT ac-tivity leads to a high-intensity dependence and viceversa (Hegerl and Juckel, 1993). Several neurophar-macological studies on P3b, the late positive ERP com-ponent elicited by task-relevant stimuli, have beenpublished also, however, with hardly any reported ef-fects by drugs affecting the 5-HT transmission, such asmethysergide or fenfluramine (Frodl-Bauch et al.,1999; Meador et al., 1989, 1995; Pritchard et al., 1987).Finally, to this date, few publications on 5-HT modu-lation of auditory change detection and MMN genera-tion have been published. This is unfortunate, becauseof the configuration of the auditory–cortical 5-HT fiberperiodicities, which has been found to overlap so-calledisofrequency bands (DeFelipe et al., 1991) associatedwith tonotopic organization of the auditory cortex,which presumably underlies MMNm/MMN generation(May et al., 1999; Tiitinen et al., 1993). This configura-tion of 5-HT fibers engenders an expectation that the5-HT projections would also play a role in detection ofpitch changes at the auditory cortex.

The brain serotonergic function can be manipulatedby acute tryptophan depletion (ATD), a dietary inter-vention that rapidly lowers plasma tryptophan, theamino acid precursor of serotonin. Previous studiesindicate that ATD also rapidly decreases tryptophanlevels (Carpenter et al., 1998) and 5-HT metaboliteconcentrations in cerebrospinal fluid (CSF) (Moja et al.,1989; Bel and Artigas, 1996; Stancampiano et al., 1997;Carpenter et al., 1998; Williams et al., 1999) and re-duces 5-HT synthesis at many cortical regions by 5 hafter ATD in humans (Nishizawa et al., 1997). Giventhe lack of information on 5-HT modulation of involun-tary attention and associated electromagnetic brainresponses, the present study aimed at elucidating therole of the 5-HT system in involuntary attention bycombining behavioral and high-resolution EEG/MEGmeasurements after ATD.

1053SEROTONERGIC MODULATION OF INVOLUNTARY ATTENTION

METHODS

Subjects and Design

Thirteen healthy, nonsmoking, paid volunteers(21–30 years old; 7 females) were included in the studyafter a medical examination, blood tests, and mentalproblem screening by the Symptoms Check List (SCL-90) questionnaire (Derogatis et al., 1973). The subjectsreported having used no drugs during the 2 weekspreceding the study; they were instructed to avoid al-cohol for at least 48 h and caffeine for 12 h before theMEG recordings. An informed written consent and in-stitutional ethical committee approval were obtained.

The study was conducted in two sessions separatedby 1 week, by using a double-blind, crossover design.The order of ATD vs control condition was randomizedacross the subjects. In each session, the subjects ar-rived at the laboratory at approximately 7:45 a.m. afteran overnight fast. After a baseline heparinized bloodsample was obtained for plasma total and free trypto-phan, and for large neutral amino acids (LNAA), thesubject was administered with either ATD or controlmixture, each being composed of 15 amino acids withor without tryptophan, as described by Young et al.(1985). During 20–30 min, the subjects first swallowedcapsules containing arginine, cysteine, and methionine(because of their very unpleasant taste) and thereafterconsumed the remaining amino acids (i.e., alanine,glycine, histidine, isoleucine, leucine, lysine, phenylal-anine, proline, serine, thereonine, tyrosine, and valine)as a chilled 300-cc strawberry/mint-flavored drink.

Previous studies indicate that the brain 5-HT syn-thesis is significantly decreased 5 h after ATD (Nishi-zawa et al., 1997); therefore, the subjects were allowedto rest in a quiet room for 5 h after the ingestion of theamino acid mixture. After the 5-h resting period, thesecond blood sample was taken for measurements ofplasma tryptophan and LNAA. Subsequently, subjec-tive levels of anxiety, mood, and fatigue were measuredwith visual analogue scales. Finally, approximately 6 hafter the onset of mixture ingestion, the subject wasguided to the shielded room for EEG and MEG mea-surements.

Blood (10 ml) was drawn from the left ulnar vein intoa vacuum tube (Venoject, Belgium), immediatelycooled, and centrifuged (2000 revolutions/min) for 10min. After centrifugation, the plasma was separatedand stored at �20°C until its assay for the amino acids.A modified procedure described by Qureshi et al. (1984)was used in the plasma assay for amino acids. Coeffi-cients of variation of 1.2, 1.01, 1.1, 1.1, 5.5, and 1.0%were determined for tryphophan, tyrosine, valine, phe-nylalanine, isoleucine, and leucine (at 150 pmol/�l),respectively. The detection limit was 10 pmol/�l.

Stimuli and Tasks

The subjects were presented with random sequencesof pure tones (10-ms rise and fall times) that equiprob-ably were either 200 or 400 ms in duration with an1110-ms offset-to-onset ISI. The tone frequency, being700 Hz for “standard” tones (p � 0.88) and either 630or 770 Hz for “deviant” tones (p � 0.06 for each), variedindependently of the duration. The stimuli were pre-sented at 60 dB over the individual hearing threshold(confirmed before the measurement) via a plastic tubeand an ear piece to the left ear, in two blocks, withresting periods of about 60 s between the blocks.

Figure 1 illustrates the procedure of the forced-choice RT task, in which the subjects were instructedto distinguish the tone duration, varying equiprobablyacross all frequencies, by pressing a button as rapidlyas possible after each tone. More specifically, the sub-jects were asked to press one button with their left-hand thumb after hearing the 200-ms tone and anotherbutton with their right-hand thumb after hearing the400-ms tone. The subjects were warned in advancethat the frequency of the tones could vary but that theywere to respond only to the tone duration. The RT andHR were measured to the standards (discounting stim-uli occurring immediately after deviants), to the devi-ants, and to the standards presented immediately afterthe deviants. To measure the distractibility by thetask-irrelevant frequency change, the RT lag to slightdeviants (RT to deviants minus RT to standards) andto standards after deviants (RT to standards after de-viants minus RT to standards) were calculated.

EEG and MEG Measurements

MEG and EEG were recorded (passband 0.03–100Hz, sampling rate 394 Hz) in a magnetically shielded

FIG. 1. Schematic illustration of the behavioral task. Stimulusduration (equiprobably either 200 or 400 ms) was to be discriminated bypressing either of two designated response buttons. Occasionally, thefrequency of the tones changed. The responses to these sounds wereclassified as “deviant” stimulus responses, and the responses to the700-Hz sounds were classified as “standard” stimulus responses. TheRT lags caused by the task-irrelevant changes in tone frequency werethen determined by subtracting the standard-sound RT from the devi-ant-sound RT. EEG/MEG deviant minus standard difference responseswere obtained analogously. Further, in the behavioral analyses, responsesto standards presented after each deviant were separately analyzed.

1054 AHVENINEN ET AL.

room (Euroshield Ltd., Eura, Finland). The subjectssat in a comfortable chair with their head placed insidea helmet-shaped whole-head MEG instrument with122 planar gradiometers (Ahonen et al., 1993) (4-DNeuroimaging Oy, Finland). Each two-channel MEGsensor unit measures two independent magnetic-fieldgradient components, �Bz/�x and �Bz/�y, the z axisbeing normal to the local helmet surface. The nose-referenced 64-channel EEG was recorded with anelectrode cap with 64 Ag/AgCl electrodes (Virtanenet al., 1996) and biopotential amplifiers specificallydesigned for simultaneous use with MEG (Virtanen etal., 1997). Bipolar vertical and horizontal electrooculo-grams (EOG) were recorded (passband 0.5–30 Hz). Thelocation of the marker coils in relation to the cardinalpoints of the head (nasion, left and right preauricularpoints) were determined with an Isotrak three-dimen-sional digitizer (Polhemus, Colchester, VT). The mag-netic fields produced by the coils were used for deter-mining the position of the subject’s head in relation tothe MEG instrument before each stimulus block.

Specific filters targeting the frequencies that encom-pass the responses of interest were used. In EEG,digital band-pass filtering was performed off-line at1–30 Hz for N1 and P2. N2 (MMN/N2b), P3a, and RONresponses to deviants were band-pass filtered at 0.5–12Hz. In MEG, N1m was filtered at 1–30 Hz and MMNmat 0.5–12 Hz. The analysis period for averaged epochswas 800 ms (including a 100-ms prestimulus baseline).The first 20 responses and all the epochs coincidingwith EOG, EEG, or MEG changes exceeding 100 �V,150 �V, or 3000 fT/cm, respectively, were omitted fromaveraging. At least 100 responses (25 for each deviantsubset) to the deviants were averaged.

Electromagnetic brain responses were averaged sep-arately for standard and deviant stimuli. The devi-ance-related responses were measured from subtrac-tion curves (the pooled deviant-stimulus responseminus standard-stimulus response). N2, P3a, and RONpeak amplitudes were determined from the 64 EEGchannels at the peak latency estimated at the FCzelectrode location, by using signal-space projection(Uusitalo and Ilmoniemi, 1997). Specifically, the spa-tial distribution for each component at the peak la-tency was estimated from the grand-averaged ERPwaveforms measured in the control condition. Subse-quently, average peak amplitude of the 64 electrodelocations, weighted by the grand-average spatial dis-tribution of a given component, was calculated sepa-rately for N2, P3a, and RON. This estimate was used todissociate the signal from the noise at the individuallevel, resulting in improved signal-to-noise ratio. Thepeak latencies and amplitudes for the ERP componentsto the standard tones were determined at the electrodelocation FCz. The deviant ERP data of three subjectswere rejected because of technical reasons.

As for the MEG waveforms, the peak amplitudes andlatencies of N1m and MMNm were quantified from thechannel pairs showing the highest-amplitude re-sponses over the left and right temporal areas. (Leftmonaural stimulation is known to elicit a MEG signalover both the right, contralateral, and the left, ipsilat-eral, auditory cortices.) The amplitudes were deter-mined as a square root of sums of squares of the am-plitudes at the same latency from the orthogonalsensor pair (a � [(�Bz/�x)2 � (�Bz/�y)2]1/2) showing thelargest amplitude response. Equivalent current dipoles(ECD) for N1m and MMNm were fitted separately forthe left and right auditory cortices based on signalsfrom a subset of 34 channels above each hemisphere. Aspherical head model with the center of symmetry at{x, y, z} � {0, 0, 45 mm} was utilized. In the coordinatesystem, the x axis points from the left to the rightpreauricular point, the y axis is perpendicular to the xaxis and passes through the nasion, and the z axispoints upward. The ECD parameters were determinedto explain the measured data optimally in the least-squares sense.

One-way analysis of variance (ANOVA) was used toanalyze the ATD vs control mixture differences in theERP peak latencies and amplitudes and in the behav-ioral variables (Statistica 4.1 software; Stat Soft Inc.,Oklahoma, USA). As for the MEG data, the N1m andMMNm peak latencies and amplitudes were tested byusing a mixture (ATD vs control mixture) by hemi-sphere ANOVA with a priori contrasts (Winer et al.,1991). A mixture by time point (baseline vs 5 h afteradministration) ANOVA with contrasts was used forthe plasma total and free tryptophan concentrationsand for the tryptophan/LNAA and LNAA/tyrosine ra-tios. For these four biochemical variables, the differ-ence between the baseline and the 5-h concentrationswas calculated and correlated with EEG, MEG, andbehavioral measures of involuntary attention by usingthe Pearson correlation coefficient.

RESULTS

The biochemical results are shown in Table 1. Nosignificant differences between ATD and control condi-tions were observed in any of the baseline concentra-tions. However, there was a highly significant mixture(ATD vs control mixture) by time point (baseline vs 5 h)interaction in ANOVA (F(1,12) � 41.3, P � 0.001) forplasma total tryptophan levels, which were reduced by75% after ATD and increased by 20% after controlmixture administration. Similar mixture by time pointinteraction (F(1,12) � 6.8, P � 0.05) was observed forthe plasma free tryptophan, indicating a 35% decreaseafter ATD and a 20% increase after control mix-ture. The tryptophan/LNAA ratio was significantly(F(1,12) � 37.1, P � 0.001) more decreased after ATD(92% decrease) than after control mixture administra-

1055SEROTONERGIC MODULATION OF INVOLUNTARY ATTENTION

tion (32% decrease). However, the ATD effects onLNAA/tyrosine were nonsignificant.

ATD significantly (F(1,10) � 12.2, P � 0.01) in-creased the depressive mood (visual analogue scalescore 40 � 21 after control mixture and 20 � 18 afterATD) in the subjects, but the levels of anxiety or vague-ness were not significantly affected. The depressivemood did not, however, correlate with the biochemicalvariables (concentration difference, baseline vs 5 h af-ter administration).

The behavioral data are presented in Table 2. Therewere no effects in the overall RT or HR for standards orin the RT lag or HR for deviants. However, in stan-dards immediately after deviants, significantly in-creased RT lag (F(1,12) � 7.2, P � 0.05) and a trendtoward HR reduction were observed after ATD (Table2). However, the correlation of these behavioral resultswith the biochemical variables was not statisticallysignificant.

The ERP results are presented in Table 3. Further,Fig. 2 indicates that the peak amplitude of N2, elicitedby the task-irrelevant deviants, was significantly re-duced by ATD (F(1,9) � 6.5, P � 0.05), but no signifi-cant effects were observed in the N2 peak latency or inP3a or RON. However, there was a trend of RONreduction after ATD. The P1, N1, and P2 peak ampli-tudes and latencies were very similar after ATD andcontrol condition.

The two-way ANOVA indicated that MMNm peak la-tency, determined from the waveforms, was significantly(F(1,12) � 10.3, P � 0.01) increased after ATD (Fig. 3,Table 4). The contrasts indicated that this effect wassignificant over both right (F(1,12) � 6.3, P � 0.05) andleft (F(1,12) � 9.1, P � 0.05) hemispheres, i.e., ipsilater-ally and contralaterally to the ear stimulated. No signif-icant effects were found in the MMNm waveform ampli-tude or in the ECD source locations or amplitudes forMMNm (Table 4). The N1m peak amplitudes and laten-cies, determined from the MEG waveforms, were verysimilar after ATD and control condition. Also, no signifi-cant effects were observed in the ECD source locations ordipole amplitudes of N1m (Table 5).

DISCUSSION

Electromagnectic responses elicited by task-irrele-vant frequency changes were significantly affected byATD. The N2 deflection, presumed to consist of theautomatic MMN and active N2b components (Naa-tanen et al., 1982; Naatanen, 1992), was significantlyreduced in amplitude. In MEG, which is sensitive totangentially oriented source components (and whichpoorly detects the N2b component), the peak latency ofMMNm was significantly increased over both hemi-spheres. Blood analyses indicated a significant reduc-tion of plasma total and free tryptophan levels after

TABLE 1

Group Mean � SD of Amino Acid Concentrations

Baseline 5 h after administration

ATD Control ATD Control

Total tryptophan (�mol/L) 63.8 � 12.7 61.9 � 27.1 N.S. 15.9 � 4.2 100.9 � 25.8***Free tryptophan (�mol/L) 5.4 � 2.3 7.1 � 3.3 N.S. 3.5 � 1.7 8.5 � 3.9**Tryptophan/LNAA 15.6 � 5.5 12.1 � 2.8 N.S. 1.3 � 0.7 8.2 � 2.1***LNAA/Tyrosine 10.7 � 1.7 10.6 � 2.8 N.S. 5.1 � 1.8 4.7 � 1.5 N.S.

Note. LNAA, large neutral amino acids; N.S., nonsignificant in contrasts.* P � 0.05.

** P � 0.01.*** P � 0.001.

TABLE 2

Mean � SD of Reaction Time (RT) andHit Rate (HR) Variables

ATD Control

RT (ms) to standards 602 � 60 602 � 43RT lag (ms) to deviants 22 � 24 15 � 20RT lag (ms) after deviants 14 � 17 2 � 14*HR (%) to standards 90 � 14 95 � 2HR (%) to deviants 89 � 14 95 � 3HR (%) after deviants 88 � 16 96 � 3

* P � 0.05.

TABLE 3

Mean � SD of the ERP Peak Amplitudes and Latencies

Peak amplitude (�V) Latency (ms)

ATD Control ATD Control

N1 �3.3 � 2.1 �3.3 � 1.8 98 � 8.0 99 � 8.8P2 1.7 � 1.6 1.6 � 2.1 173 � 11.0 174 � 15.4N2 �1.6 � 1.0 �2.4 � 0.6* 183 � 34.9 179 � 30.1P3a 3.9 � 2.0 3.7 � 2.4 328 � 20.5 324 � 30.8RON �1.2 � 2.0 �2.1 � 1.57 510 � 72.2 511 � 20.2

* P � 0.05.

1056 AHVENINEN ET AL.

ATD in the subjects. Previous studies have shown thatsuch a reduction of plasma tryptophan concentrationrapidly decreases central 5-HT metabolite concentra-tions in human CSF (Carpenter et al., 1998). Positronemission tomography (PET) studies have further indi-

cated that the brain 5-HT synthesis is significantlydiminished by 5 h after ATD in humans (Nishizawa etal., 1997). Therefore, the present ERP and MEG re-sults may be related to changes in the brain 5-HTfunction after ATD.

FIG. 2. (a) Grand-average EEG difference waves (deviant minus standard ERP) measured during ATD and under control conditions. (b)Grand-average EEG potential maps on a spherical head model, estimated at the peak latency of N2, including MMN and N2b subcompo-nents, under each condition.

FIG. 3. (a) MEG subtraction curves (frequency deviants minus standards) in a representative subject. (b) Right-hemispheric equivalentcurrent dipole fitted for MMNm in the same subject and the 10-fT isocontour lines of the magnetic field.

TABLE 4

Mean � SD Peak Amplitudes and Latencies of N1m and MMNm, Elicited to Left-Ear Tones

Peak amplitude (fT/cm) Latency (ms)

ATD Control ATD Control

N1m Contralateral 47 � 20 51 � 14.5 91 � 20 99 � 17Ipsilateral 35 � 14 32 � 16.6 101 � 16 104 � 9

MMNm Contralateral 44 � 16 50 � 15.0 185 � 49 162 � 33*Ipsilateral 27 � 10 30 � 11.2 176 � 25 157 � 24*

* P � 0.05.

1057SEROTONERGIC MODULATION OF INVOLUNTARY ATTENTION

MMN and MMNm to Task-Irrelevant Sound Changes

Supporting previous observations (Hari et al., 1984),the MEG dipole modeling results suggested that theMMNm, supposed to be a magnetic counterpart of thesupratemporal MMN subcomponent (Naatanen, 1992),is generated at the temporal lobe, near the primaryauditory cortices. PET studies have shown that ATDsignificantly reduces 5-HT synthesis at the vicinity ofthis brain region in humans (Nishizawa et al., 1997).Moreover, a large amount of 5-HT immunoreactivefibers has been found at the primate auditory cortex(Campbell et al., 1987), and cat studies suggest thatthere are 5-HT fibers configured into periodicities thatoverlap auditory-cortical isofrequency bands (DeFelipeet al., 1991). The 5-HT neurons have, furthermore,been shown to interact with GABA neurons (DeFelipeet al., 1991), crucial for frequency tuning of tonotopi-cally organized cortical auditory neurons (Wang et al.,2000) and for the MMN generation (Javitt et al., 1996).Hence, decreased 5-HT could flatten the tuning curveof tonotopically organized neurons, which have beensuggested to underlie MMNm/MMN generation in theauditory cortices (May et al., 1999; Tiitinen et al.,1993). The decreased coherency of tonotopic organiza-tion in the auditory cortex could, in turn, result indecreased MMN amplitude and delayed MMNm peaklatency. However, the exact underlying neurophysio-logical mechansims remain to be determined in furtherempirical efforts.

Previous studies have indicated that MMN (Woldorffet al., 1991) and MMNm (Woldorff et al., 1998) gener-ation are attenuated when the subject focuses on a taskwith high attentional demand. Based on these results,the present MMNm/N2 effects could also be inter-preted to reflect improved ability to allocate attentionalresources to the duration-discrimination task afterATD. This line of reasoning can be supported by arecent neuropsychological result showing that the sub-jects’ ability to focus attention, both during dichoticlistening and in the Stroop test (Stroop, 1935), wassignificantly improved after ATD (Schmitt et al., 2000).As also proposed by Schmitt et al. (2000), such changes

in focused attention could be related to decrease of5-HT release in the prefrontal brain regions, which hasbeen previously shown in invasive rat studies (Bel andArtigas, 1996) and PET measurements in humans(Nishizawa et al., 1997). More specifically, 5-HT neu-rons have been shown to inhibit prefrontal dopaminefunction (Schmidt and Fadayel, 1995), and since dopa-mine probably enhances attention (Puumala andSirvio, 1998), release of the 5-HT inhibition of dopa-mine neurons could improve resource allocation to thetask-relevant activity. This assumption can be sup-ported by the finding that haloperidol, an antagonist ofdopamine D2 receptors, produced clearly opposite ef-fects: MMN to task-irrelevant deviants was increasedin amplitude, and the processing negativity, indexingselective attention to task-relevant stimulation (Naa-tanen, 1992), was decreased during dichotic listeningby decreased dopamine function (Kahkonen et al., 2001).

Conversely, the present N2 decrease and MMNmlatency delay could also reflect decreased distraction bytask-irrelevant processing. According to the prevailingtheory (Naatanen, 1992), automatic sound change de-tection in the auditory cortex, reflected by the supra-temporal MMN subcomponent and by MMNm, is fol-lowed by involuntary initiation of attention shifting,which generates the frontal MMN subcomponent.Against this theoretical background (Naatanen, 1992),the MMNm peak latency increase suggests delayedautomatic change detection in the auditory cortex byATD. However, since the MMNm amplitude was notaffected, the N2 reduction may also have been partiallyaccounted for by effects on the frontal MMN subcom-ponent, which is poorly detected by MEG. Given thatfrontal brain lesions have been shown to reduce MMN(Alho et al., 1994), N2 reduction could partially followfrom temporarily repressed activation of the frontalcortex after ATD. This theoretical perspective couldalso be used to explain the above-mentioned neuropsy-chological results of Schmitt et al. (2000): Improvedperformance in the Stroop and dichotic-listening tasksafter ATD could as well follow from reduced distracti-bility task-irrelevant processing (e.g., by attenuated

TABLE 5

Mean � SD of the MEG Source Modeling Results

Hemisphere Conditionx

(mm)y

(mm)z

(mm)Q

(nAm)Goodness of fit

(%)

N1m Contralateral ATD 51 � 8 7 � 9 56 � 10 21 � 9 90 � 8Control 49 � 7 8 � 12 62 � 13 20 � 9 88 � 11

Ipsilateral ATD �50 � 8 1 � 11 61 � 17 14 � 8 88 � 5Control �50 � 8 3 � 10 64 � 17 15 � 10 84 � 9

MMNm Contralateral ATD 47 � 13 15 � 19 57 � 15 27 � 16 81 � 8Control 46 � 10 20 � 14 62 � 13 21 � 10 82 � 5

Ipsilateral ATD �42 � 10 5 � 19 58 � 25 25 � 15 77 � 6Control �52 � 16 10 � 19 62 � 17 18 � 16 70 � 11

1058 AHVENINEN ET AL.

automatic activation of lexical processing in the Strooptask).

P3 to Task-Irrelevant Sound Changes

Supporting previous observations, a clear positivedeflection with a peak latency at around 300 ms post-stimulus was elicited by the task-irrelevant frequencychange (Schroger and Wolff, 1998a,b). According to thepresent theories (for a review, see Naatanen, 1992),this P3 to task-irrelevant stimulus is mainly domi-nated by the P3a component. Previous neuropharma-cological studies have shown that P3a is suppressed byboth ethanol (Jaaskelainen et al., 1999) and the dopa-mine D2-receptor antagonist haloperidol (Kahkonen etal., 2002), whereas muscarinic agonists may enhancethe P3a counterpart in monkeys (O’Neill et al., 2000).As for the 5-HT system, few prior efforts have beenpublished, but since no significant ATD effect on the P3to task-irrelevant tones were observed here, one mightsuggest that this response might not be modulated bythe 5-HT system. Perhaps the role of 5-HT modulationis more prevalent in the earlier, modality-specific,phases of involuntary attention that are thought to berepresented by the subcomponents of MMN (Naatanen,1992). However, this does not mean that the MMN/N2bcomponents should necessarily be more sensitive mea-sures of involuntary attention than P3a per se.

It has to be noted that the task-irrelevant frequencychange (70 Hz decrease or increase) was multiply re-peated during the stimulus block; in other words, thischange was not a “genuinely” novel stimulus. Addition-ally, the paradigm was actually geared not towardeliciting robust P3a responses, but rather toward min-imizing the N1/N2b contamination of the MMN re-sponse. Thus, further P3a studies utilizing novel tonesembedded in stimulus blocks would be informative.Another factor complicating the interpretation of thepresent P3a results could be related to possible overlapby other P3 subcomponents, such as P3b. This is sug-gested, for instance, by the relatively posterior grand-average scalp distribution of the P3 to task-irrelevantsound changes (Fig. 2) and by its peak latency, whichsomewhat resembles that of the P3b. However, the factthat the P3 to task-irrelevant sounds was determinedfrom the subtraction curves reduces the possibility thatsuch P3b overlap would be associated with the task-relevant activity. Furthermore, little evidence of 5-HTmodulation of the P3b component has been found thusfar (see Frodl-Bauch et al., 1999), and it is thereforeimprobable that the (possibly) overlapping P3b couldhave altered P3a results with respect to 5-HT.

MEG and EEG Responses to Standard Tones

No ATD effects were observed in N1, P2, and N1m.This is generally in line with the observation that theserotonin-depleting agents, such as methysergide

(Meador et al., 1989) or fenfluramine (Meador et al.,1995), do not significantly affect N1 or P2 amplitude orpeak latency. Thus, it appears that the 5-HT systemdoes not modulate the generation of ERP and MEGresponses elicited by constantly repeated standardtones at the 100- to 200-ms latency range. However, ithas to be noted that there is some evidence that theintensity dependence of N1/P2 components could beregulated by 5-HT (Hegerl and Juckel, 1993).

Behavioral Effects

The behavioral distractibility, i.e., RT lag to fre-quency deviants, was not significantly affected by ATD.One explanation is that EEG and MEG are more sen-sitive in the detection of impairments in 5-HT modu-lation of involuntary attention. This could be based, forexample, on a better signal-to-noise ratio in electro-physiological vs behavioral measures. However, it hasto be noted that this interpretation might also be com-plicated by the adaptive changes (down-regulationand/or up-regulation) of receptors after acute drugchallenge or a comparable manipulation such as ATD.A recent PET study suggests adaptive changes in5-HT2 receptors subtype after ADT, which was inter-preted to be an adaptive change to acutely decreased5-HT in the brain (Yatham et al., 2001). Therefore achronic, and more severe, deficit in the brain 5-HTfunction (e.g., in specific patient groups) might producemore pronounced attention effects. Furthermore, de-spite the fact that the RT lag to task-irrelevant toneswas not affected by ATD, the RT lag to standardsimmediately after task-irrelevant frequency changewas significantly increased. There was also a trendtoward impaired HR to these stimuli. Thus, it appearsthat ATD impaired the subjects’ ability to shift back tothe relevant task after distraction. In addition, therewas a trend toward reduction of RON that hasbeen presumed to index reorienting after distraction(Schroger and Wolff, 1998a).

Possible Limitations

ATD appeared to increase the subjective feelings ofdepression. However, this result was observed only inthe visual analogue scale and was not confirmed withmore objective means of mood assessment. Further-more, it is important to note that the present resultswere mainly observed in physiological measures of in-voluntary attention, whereas the task-relevant perfor-mance (simple RT and HR in duration discrimination)was not affected at all.

Finally, a possible limitation to the present conclu-sions may stem from confounding effects of other trans-mitter systems. For instance, based on depletion stud-ies of other amino acids (Harmer et al., 2001), possibletyrosine deficiency might have interfered with centralcatecholamine function. However, the results indicated

1059SEROTONERGIC MODULATION OF INVOLUNTARY ATTENTION

that LNAA/tyrosine ratio was not significantly affectedby ATD, showing that metabolism of dopamine and nor-adrenaline was not affected in the present experiment.

CONCLUSION

The MEG and EEG results suggest that the reduc-tion of the brain 5-HT function by ATD may delayautomatic change detection at the auditory cortex andthus impair initiation of involuntary attention shiftingreflected by the N2 component. This process could alsohave been partially overlapped by changes in 5-HTmodulation of attentional resource allocation to thetask-relevant activity. Future studies combining EEG,MEG, and behavioral measures could help us to en-hance the understanding of the neurochemical basis ofvarious attentional disorders.

ACKNOWLEDGMENTS

This work was supported by the Helsinki University Central Hos-pital Research Funds, the Finnish Cultural Foundation, and theAcademy of Finland. We thank Ms. Suvi Heikkila and Mr. TeemuPeltonen for their assistance in carrying out the experiments.

REFERENCES

Ahonen, A., Hamalainen, M., Kajola, M., Knuutila, J., Laine, P.,Lounasmaa, O., Parkkonen, L., Simola, J., and Tesche, C. 1993.122-channel SQUID instrument for investigating the magneticsignals from the human brain. Physica Scripta T49: 198–205.

Ahveninen, J., Jaaskelainen, I. P., Pekkonen, E., Hallberg, A.,Hietanen, M., Naatanen, R., Schroger, E., and Sillanaukee, P.2000. Increased distractibility by task-irrelevant sound changes inabstinent alcoholics. Alcohol. Clin. Exp. Res. 24: 1850–1854.

Alho, K., Woods, D. L., Algazi, A., Knight, R. T., and Naatanen, R.1994. Lesions of frontal cortex diminish the auditory mismatchnegativity. Electroencephalogr. Clin. Neurophysiol. 91: 353–362.

Barkley, R. A. 1998. Attention-Deficit Hyperactivity Disorder: A Hand-book for Diagnosis and Treatment, 2nd ed. Guilford, New York.

Bel, N., and Artigas, F. 1996. Reduction of serotonergic function inrat brain by tryptophan depletion: Effects in control and fluvox-amine-treated rats. J. Neurochem. 67: 669–676.

Berchou, R., Chayasirisobhon, S., Green, V., and Mason, K. 1986.The pharmacodynamic properties of lorazepam and methylpheni-date drugs on event-related potentials and power spectral analysisin normal subjects. Clin. Electroencephalogr. 17: 176–180.

Campbell, M. J., Lewis, D. A., Foote, S. L., and Morrison, J. H. 1987.Distribution of choline acetyltransferase-, serotonin-, dopamine-beta-hydroxylase-, tyrosine hydroxylase-immunoreactive fibers inmonkey primary auditory cortex, J. Comp. Neurol. 261: 209–220.

Carpenter, L. L., Anderson, G. M., Pelton, G. H., Gudin, J. A.,Kirwin, P. D. S., Price, L. H., Heninger, G. R., and McDougle, C. J.1998. Tryptophan depletion during continuous CSF sampling inhealthy human subjects Neuropsychopharmacology 19: 26–35.

DeFelipe, J., Hendry, S. H., Hashikawa, T., and Jones, E. G. 1991.Synaptic relationships of serotonin-immunoreactive terminal bas-kets on GABA neurons in the cat auditory cortex. Cerebral Cortex1: 117–133.

Derogatis, L. R., Lipman, R. S., and Covi, L. 1973. SCL-90: Anoutpatient psychiatric rating scale—preliminary report. Psycho-pharmacol. Bull. 9: 13–28.

Dolan, R. J., Bench, C. J., Brown, R. G., Scott, L. C., and Frackowiak,R. S. 1994. Neuropsychological dysfunction in depression: Therelationship to regional cerebral blood flow. Psychol. Med. 24:849–857.

Duncan, C. C., and Kaye, W. H. 1987. Effects of clonidine on event-related potential measures of information processing. Electroen-cephalogr. Clin. Neurophysiol. Suppl. 40: 527–531.

Friedman, D., Cycowicz, Y. M., and Gaeta, H. 2001. The novelty P3:An event-related brain potential (ERP) sign of the brain’s evalua-tion of novelty. Neurosci. Biobehav. Rev. 25: 355–373.

Frodl-Bauch, T., Bottlender, R., and Hegerl, U. 1999. Neurochemicalsubstrates and neuroanatomical generators of the event-relatedP300. Neuropsychobiology 40: 86–94.

Hari, R., Hamalainen, M., Ilmoniemi, R. J., Kaukoranta, E., Reini-kainen, K., Salminen, J., Alho, K., Naatanen, R., and Sams, M.1984. Responses of the primary auditory cortex to pitch changes ina sequence of tone pips: Neuromagnetic recordings in man. Neu-rosci. Lett. 50: 127–132.

Harmer, C. J., McTavish, S. F., Clark, L., Goodwin, G. M., andCowen, P. J. 2001. Tyrosine depletion attenuates dopamine func-tion in healthy volunteers. Psychopharmacology 154: 105–111.

Harris, R. A. 1999. Ethanol actions on multiple ion channels: Whichare important? Alcohol. Clin. Exp. Res. 23: 1563–70.

Hegerl, U., and Juckel, G. 1993. Intensity dependence of auditoryevoked potentials as an indicator of central serotonergic neuro-transmission: A new hypothesis. Biol. Psychiatry 33: 173–187.

Hirvonen, J., Jaaskelainen, I., Saher, M., Pekkonen, E., Sillanaukee,P., and Naatanen, R. 1998. Temazepam and ethanol effects onMMN. Paper presented at the I International Workshop on MMNand Its Clinical Applications, Helsinki.

Jaaskelainen, I. P., Naatanen, R., and Sillanaukee, P. 1996. Effect ofacute ethanol on auditory and visual event-related potentials: Areview and reinterpretation. Biol. Psychiatry 40: 284–291.

Jaaskelainen, I. P., Schroger, E., and Naatanen, R. 1999. Electro-physiological indices of acute effects of ethanol on involuntaryattention shifting. Psychopharmacology 141: 16–21.

Javitt, D. C., Steinschneider, M., Schroeder, C. E., and Arezzo, J. C.1996. Role of cortical N-methyl-D-aspartate receptors in auditorysensory memory and mismatch negativity generation: Implica-tions for schizophrenia. Proc. Natl. Acad. Sci. USA 93: 11962–11967.

Kahkonen, S., Ahveninen, J., Jaaskelainen, I. P., Kaakkola, S., Naa-tanen, R., Huttunen, J., and Pekkonen, E. 2001. Effects of halo-peridol on selective attention—A combined whole-head MEG andhigh-resolution EEG study. Neuropsychopharmacology 25: 498–504.

Kahkonen, S., Ahveninen, J., Pekkonen, E., Kaakkola, S., Huttunen,J., Ilmoniemi, R. J., and Jaaskelainen, I. P. 2002. Dopamine D2-receptor antagonist haloperiodol impairs involuntary attentionshifting and reorienting in healthy subjects: Evidence from com-bined MEG and EEG study.

Kennedy, S. H., Evans, K. R., Kruger, S., Mayberg, H. S., Meyer,J. H., McCann, S., Arifuzzman, A. I., Houle, S., and Vaccarino, F. J.2001. Changes in regional brain glucose metabolism measuredwith positron emission tomography after paroxetine treatment ofmajor depression. Am. J. Psychiatry 158: 899–905.

Liotti, M., and Mayberg, H. S. 2001. The role of functional neuroim-aging in the neuropsychology of depression. J. Clin. Exp. Neuro-psychol. 23: 121–136.

May, P., Tiitinen, H., Ilmoniemi, R. J., Nyman, G., Taylor, J. G., andNaatanen, R. 1999. Frequency change detection in human audi-tory cortex. J. Comput. Neurosci. 6: 99–120.

McCann, U. D., Mertl, M., Eligulashvili, V., and Ricaurte, G. A. 1999.Cognitive performance in (�/�) 3,4-methylenedioxymethamphet-

1060 AHVENINEN ET AL.

amine (MDMA, “ecstasy”) users: A controlled study. Psychophar-macologia 143: 417–425.

Meador, K. J., Loring, D. W., Davis, H. C., Sethi, K. D., Patel, B. R.,Adams, R. J., and Hammond, E. J. 1989. Cholinergic and seroto-nergic effects on the P3 potential and recent memory. J. Clin. Exp.Neuropsychol. 11: 252–260.

Meador, K. J., Loring, D. W., Hendrix, N., Nichols, M. E., Oberzan, R.,and Moore, E. E. 1995. Synergistic anticholinergic and antisero-tonergic effects in humans. J. Clin. Exp. Neuropsychol. 17: 611–621.

Mervaala, E., Alhainen, K., Helkala, E. L., Partanen, J., Jousmaki, V.,Vayrynen, M., Heinonen, E., and Riekkinen, P. 1993. Electrophysio-logical and neuropsychological effects of a central alpha 2-antagonistatipamezole in healthy volunteers. Behav. Brain Res. 55: 85–91.

Moja, E. A., Cipolla, P., Castoldi, D., and Tofanetti, O. 1989. Dose-response decrease in plasma tryptophan and in brain tryptophanand serotonin after tryptophan-free amino acid mixtures in rats.Life Sci. 44: 971–976.

Moore, P., Landolt, H. P., Seifritz, E., Clark, C., Bhatti, T., Kelsoe, J.,Rapaport, M., and Gillin, J. C. 2000. Clinical and physiologicalconsequences of rapid tryptophan depletion. Neuropsychopharma-cology 23: 601–622.

Naatanen, R. 1992. Attention and Brain Function. Erlbaum, Hills-dale, NJ.

Naatanen, R., Simpson, M., and Loveless, N. E. 1982. Stimulusdeviance and evoked potentials. Biol. Psychol. 14: 53–98.

Nathan, P. J., Sitaram, G., Stough, C., Silberstein, R. B., and Sali, A.2000. Serotonin, noradrenaline and cognitive function: A prelimi-nary investigation of the acute pharmacodynamic effects of a se-rotonin versus a serotonin and noradrenaline reuptake inhibitor.Behav. Pharmacol. 11: 639–642.

Nishizawa, S., Benkelfat, C., Young, S. N., Leyton, M., Mzengeza, S.,de Montigny, C., Blier, P., and Diksic, M. 1997. Differences be-tween males and females in rates of serotonin synthesis in humanbrain. Proc. Natl. Acad. Sci. USA 94: 5308–5313.

Oades, R. D. 2000. Differential measures of ‘sustained attention’ inchildren with attention-deficit/hyperactivity or tic disorders: Rela-tions to monoamine metabolism. Psychiatry Res. 93: 165–178.

O’Neill, J., Halgren, E., Marinkovic, K., Siembieda, D., Refai, D.,Fitten, L. J., Perryman, K., and Fisher, A. 2000. Effects of musca-rinic and adrenergic agonism on auditory P300 in the macaque.Physiol. Behav. 70: 163–170.

Pekkonen, E., Hirvonen, J., Jaaskelainen, I. P., Kaakkola, S., andHuttunen, J. 2001. Auditory sensory memory and the cholinergicsystem: Implications for Alzheimer’s disease. NeuroImage 14:376–82.

Potts, G. F., Dien, J., Hartry-Speiser, A. L., McDougal, L. M., andTucker, D. M. 1998. Dense sensor array topography of the event-related potential to task-relevant auditory stimuli. Electroen-cephalogr. Clin. Neurophysiol. 106: 444–456.

Pritchard, W. S., Raz, N., and August, G. J. 1987. No effect of chronicfenfluramine on the P300 component of the event-related poten-tial. Int. J. Neurosci. 35: 105–110.

Puumala, T., and Sirvio, J. 1998. Changes in activities of dopamineand serotonin systems in the frontal cortex underlie poor choiceaccuracy and impulsivity of rats in an attention task. Neuroscience83: 489–499.

Qureshi, G. A., Fohlin, L., and Bergstrom, J. 1984. Application ofhigh-performance liquid chromatography to the determination offree amino acids in physiological fluids. J. Chromatogr. 297: 91–100.

Ricaurte, G. A., Martello, A. L., Katz, J. L., and Martello, M. B. 1992.Lasting effects of (�/�)-3,4-methylenedioxymethamphetamine(MDMA) on central serotonergic neurons in nonhuman primates:Neurochemical observations. J. Pharmacol. Exp. Therap. 261:616–622.

Robbins, T. W., Jones, G. H., and Sahakian, B. J. 1989. Centralstimulants, transmitters and attentional disorder: A perspectivefrom animal studies. In Attention Deficit Disorder: Clinical andBasic Research (T. Sagvolden and T. Archer, Eds.), pp. 199–222.Erlbaum, Hillsdale, NJ.

Schmidt, C. J., and Fadayel, G. M. 1995. The selective 5-HT2Areceptor antagonist, MDL 100,907, increases dopamine efflux inthe prefrontal cortex of the rat. Eur. J. Pharmacol. 273: 273–279.

Schmitt, J. A., Jorissen, B. L., Sobczak, S., van Boxtel, M. P., Hog-ervorst, E., Deutz, N. E., and Riedel, W. J. 2000. Tryptophandepletion impairs memory consolidation but improves focussedattention in healthy young volunteers. J. Psychopharmacol. 14:21–29.

Schroger, E., and Wolff, C. 1998a. Attentional orienting and reori-enting is indicated by human event-related brain potentials. Neu-roreport 9: 3355–3358.

Schroger, E., and Wolff, C. 1998b. Behavioral and electrophysiolog-ical effects of task-irrelevant sound change: A new distractionparadigm. Cogn. Brain Res. 7: 71–87.

Stancampiano, R., Melis, F., Sarais, L., Cocco, S., Cugusi, C., andFadda, F. 1997. Acute administration of a tryptophan-free aminoacid mixture decreases 5-HT release in rat hippocampus in vivo.Am. J. Physiol. 272: R991–R994.

Stroop, J. 1935. Studies of interference in serial verbal reactions. J.Exp. Psychol. 18: 643–662.

Teo, R. K., and Ferguson, D. A. 1986. The acute effects of ethanol onauditory event-related potentials. Psychopharmacology 90: 179–184.

Tiitinen, H., Alho, K., Huotilainen, M., Ilmoniemi, R. J., Simola, J.,and Naatanen, R. 1993. Tonotopic auditory cortex and the magne-toencephalographic (MEG) equivalent of the mismatch negativity.Psychophysiology 30: 537–540.

Uusitalo, M. A., and Ilmoniemi, R. J. 1997. Signal-space projectionmethod for separating MEG or EEG into components. Med. Biol.Eng. Comput. 35: 135–140.

Virtanen, J., Parkkonen, L., Ilmoniemi, R. J., Pekkonen, E., andNaatanen, R. 1997. Biopotential amplifier for simultaneous oper-ation with biomagnetic instruments. Med. Biol. Eng. Comput. 35:402–408.

Virtanen, J., Rinne, T., Ilmoniemi, R. J., and Naatanen, R. 1996.MEG-compatible multichannel EEG electrode array. Electroen-cephalogr. Clin. Neurophysiol. 99: 568–570.

Wang, J., Caspary, D., and Salvi, R. J. 2000. GABA-A antagonistcauses dramatic expansion of tuning in primary auditory cortex.Neuroreport 11: 1137–1140.

Williams, W. A., Shoaf, S. E., Hommer, D., Rawlings, R., and Lin-noila, M. 1999. Effects of acute tryptophan depletion on plasmaand cerebrospinal fluid tryptophan and 5-hydroxyindolacetic acidin normal volunteers. J. Neurochem. 72: 1641–1647.

Winer, B., Brown, D., and Michels, K. 1991. Statistical Principles inExperimental Design. McGraw–Hill, New York.

Woldorff, M. G., Hackley, S. A., and Hillyard, S. A. 1991. The effectsof channel-selective attention on the mismatch negativity waveelicited by deviant tones. Psychophysiology 28: 30–42.

Woldorff, M. G., Hillyard, S. A., Gallen, C. C., Hampson, S. R., andBloom, F. E. 1998. Magnetoencephalographic recordings demon-strate attentional modulation of mismatch-related neural activityin human auditory cortex. Psychophysiology 35: 283–292.

Yatham, L. N., Liddle, P. F., Shiah, I. S., Lam, R. W., Adam, M. J.,Zis, A. P., and Ruth, T. J. 2001. Effects of rapid tryptophan deple-tion on brain 5-HT(2) receptors: A PET study. Br. J. Psychiatry178: 448–453.

Young, S. N., Smith, S. E., Pihl, R. O., and Ervin, F. R. 1985.Tryptophan depletion causes a rapid lowering of mood in normalmales. Psychopharmacology 87: 173–177.

1061SEROTONERGIC MODULATION OF INVOLUNTARY ATTENTION