- Email: [email protected]
Neuromagnetic responses of human auditory cortex to interruptions in a steady rhythm
164
Neuroscience Letters, 99 (1989) 164 168 Elsevier ScientificPublishers Ireland Ltd.
NSL 06009
Neuromagnetic responses of human auditory cortex to interruptions in a steady rhythm R. Hari, S.L. Joutsiniemi, M. HS.mfilfiinen and V. Vilkman Low Temperature Laboratory, Helsinki University ~[' Technology, Espoo (Finland) (Received 17 October 1988; Revised version received 19 December 1988; Accepted 21 December 1988) Key wordsv Auditory cortex: Rhythm; Magnetoencephalography; Evoked response; Man We have recorded, with a 7-channel SQUID gradiometer, evoked magnetic responses of 6 healthy humans to interruptions of a steady rhythm of 50 ms 'standard' tone bursts repeated once every 610 ms. Ten percent of the tones occurred 'too early', 410 ms after the precedingstimulus. The response to standards peaked, on average, at 90 ms and that to the early tones at 148 ms. Field patterns were dipolar during both responses and the equivalent sources agreed with activation of the supratemporal auditory cortex, at slightly different locations. The dipole moments were more than twice as strong for the early tones as tbr the standards. The results emphasize the importance of temporal stimulation patterns in activating the human auditory cortex,
Perception of rhythmical passages is essential for proper i n t e r p r e t a t i o n o f speech, music a n d other acoustic messages. A l t h o u g h cortical m e c h a n i s m s are certainly involved in coding of complex acoustic patterns, no d a t a exist a b o u t the reactions of the h u m a n a u d i t o r y cortex to deviations in timing of s o u n d stimuli. The cerebral magnetic fields are t h o u g h t to be m a i n l y generated by neural currents in the fissural cortex a n d the technique is, therefore, well suited for studies o f cortical areas e m b e d d e d in the Sylvian fissure. Earlier recordings have revealed how the evoked activity of the s u p r a t e m p o r a l a u d i t o r y cortex changes in time a n d location a n d how some characteristics of acoustic stimuli are represented in cortical f u n c t i o n a l feature maps (for a review, see ref. 4). Rare deviations of physical stimulus features, like frequency or intensity, in an otherwise m o n o t o n o u s sequence of ' s t a n d a r d ' tones, have been f o u n d to elicit an extra magnetic response, ' m i s m a t c h field' at the a u d i t o r y cortex [4, 11]. In the present work we show that i n t e r r u p t i o n s in a steady r h y t h m of identical stimuli m a y also change the activity of the a u d i t o r y cortex. Seven healthy adults (3 females, 4 males) were studied in our magnetically shielded room. D u r i n g the experiment the subject was lying o n a bed with his head supported by a v a c u u m cast. The s t i m u l a t i o n resembled that used in earlier electric evoked Correspondence: R. Hari, Low Temperature Laboratory, Helsinki University of Technology, SF-02150 Espoo, Finland. 0304-3940/89/$ 03.50 ':~71989 ElsevierScientificPublishers Ireland Ltd,
165
potential studies to too-early presentation of somatosensory, auditory and visual stimuli [3, 8, 10]. Sinusoidal 1 kHz tone bursts (rectangular envelope, duration 50 ms, intensity 78 dB SPL) were presented to the left ear through a plastic tube and earpiece. The steady rhythm of 'standard' tones having an interstimulus interval (ISI) of 610 ms was randomly interrupted by an early tone 410 ms after the previous standard. Both stimuli were physically identical and the ISI following the early tone was always 610 ms. Early tones consisted 10% of all stimuli; the number of standards after each early tone varied randomly with an even distribution between 3 and 15. Subjects either ignored all stimuli by concentrating on reading a novel or paid attention to the tones by counting the number of early stimuli. The magnetic field over the right hemisphere was measured with a 7-sensor lstorder gradiometer S Q U I D (Superconducting QUantum Interference Device) system of high sensitivity (5 - 6 t T / x / ~ intrinsic noise level) [9]. The pickup coils, separated by 36.5 mm, form a hexagonal array on a spherical surface (radius 125 mm), 18-20 mm above the scalp. The bandpass-filtered signals (3 dB points at 0.05 and 250 Hz, roll-off for the highpass filter 35 dB/decade and more than 80 dB/decade for the low-pass filter) were digitized at 1200 or 1500 Hz, and 600-900 artefact-free responses for standards and 70-100 for deviants were averaged on-line. The responses to the first standard after each deviant were not included in the average. In all experiments, the vertical electrooculogram was recorded and magnetic responses coinciding with blinks or eye movements were discarded from the average. S.E.M.s of the averaged responses were used as estimates for the total experimental noise, needed for confidence limit calculations. Typical values were 6-10 ffF for standards and 15-30 IT for the early tones. The signal channel with maximal amplitudes was chosen for further analysis in all subjects. The amplitudes were measured with respect to a 40 ms prestimulus baseline. Field maps were obtained from four subjects (S1, $3, $4, and $7) by measuring responses at 35-49 sensor locations over the right hemisphere. Equivalent current dipoles were found using a spherically symmetric conductor model with a center of symmetry estimated from the shape of the head in the measurement area; in this case the local radius of curvature was 12 cm. The fitting algorithm used to compute the equivalent dipoles took into account the exact sensor locations and orientations, measured with respect to the head [9], so that the effect of source and volume currents could be taken properly into account. Fig. 1 shows responses of two subjects over the right hemisphere in the attend condition. The responses of S1 to standards peak at 75 ms, and resemble the Nl00mdeflections reported earlier to different types of auditory stimuli [4, 5]. The responses to the early stimuli are clearly higher in amplitude than those to standards (360 vs 150 fF) and peak at 100-105 ms. Similar responses were obtained during repeated sessions. The spatial distribution of responses does not clearly differ in this subject for standard and early stimuli. In $2, the responses to the early stimuli were more complex in waveform containing two peaks 120-145 ms after stimulus onset. Polarity reversal between channels 1 and 4, for example, suggests that the measurement has
166
Subject
@
Subiect 2
/
!9
L in --
n 2 0 0 nl'~
l]
Fig. I. Responses to changes in the steady rhythm for two subjects; the approximate measurement locations over the righl hemisphere are shown in the inserts. The recording of S1 was repeated 4 times at the same location averaging 19 39 responses to the early tones (thin lines) and 150 300 to standards (thick lines). For $2 the corresponding numbers are 53 and 500. Passband 0.05 45 Hz.
been made approximately over the source. Further, the different distributions of the first and second peak indicate activation of more than one source. In all subjects the response waveforms are clearly different for standards and early stimuli (Fig. 2). During the ignore condition the mean (_+ S.E.M.) latency was 90 (+_ 5) ms for N l 0 0 m to standards and 148 (_+ 18) ms for the response to the early tones. In 5 out of 7 subjects the response to the early stimulus was larger during the attend than ignore condition, in two other subjects the amplitudes were the same within experimental noise. During attention the mean amplitude of all subjects was SUbleCt ]
Subject 2
Subject 3
Subject 4
Subject 5
Subiect 6
qubject 7
[2,! /*,',,
,~,
,..:.,.,:.,i,,
L,
f] z
F
,,,,,
; ~ ,'v,
I 200 tT
L
,'"i,
< F - - 1
~1
I - - I
F - - I
~
l
r ~ l
200 ms
Fig. 2. Responses of all subjects in ignore and attend conditions. The continuous lines indicate responses to standards and the dotted lines those to the early stimuli, The recordings are from the right frontotemporal area, over the anterior field extremum.
167
48% larger than during the ignore condition (P<0.05, t-test). In control recordings much stronger responses were evoked by decreases than increases of the interstimulus interval. Fig. 3 shows isocontour field maps for two subjects during N100m to standards and the response to early stimuli, both in the attend condition; maps are shown at time moments with the best goodness-of-fit values for the dipole model. The locations of the equivalent dipoles agree with activation of the supratemporal auditory cortex; the dipole model explained 85 97% of the field variance in the four subjects. The 3-dimensional locations of the dipoles for the two deflections differed by 6-29 mm in the four subjects without any systematic features; the location differences were statistically significant at the 95% confidence level in 3 subjects. The dipole moments were, on average, 120% stronger for the early stimuli (mean 27 nA.m) than for the standards (mean 12 nA.m).
Subject 1 110 ms .:.
Subject 4
fY
k
..,-
. "\",.
I..,-
-,-~
~y
,,,f
....... - ,-', . . . . .
a<
;
l
/,
.:~r ,b' t,,-,'/"
Z
t
"'"
• ........
[.,-,
.°
X
". I
......
.
...';
o,•"
145ms 9 "
\'-
• "',,
,,.......... i~
. . i I I
<
/ '
,
I ;
,1
°"
. %
i
.
.
........... •
i I
.
i i L,gV//,~
°
0 r I o
,,,'.,//
o-o°°
i]ll.VJ/
.° ~" I . . o " o• •
.o"
"
"
s
•
.-o-
~
-~.
I I1~ F o-~ lellll IBa •
"
I/I:
• •
/: : : m ~ . ' . ' i ~ l . l l
'11¢ ~1 I
1/o1~
,,'#:,. -" ,: llj
•
,,
P
~
I IIIII I IIIII
,. ~ / " •
•
am-,
.--'./
; i
"-. ~ ...--'.--~
A\\-
.?:,:;12. j
..,, ". •
f'
o'°
,s
I'.--:i.~%~,'~ ,. ~ : . : ~ / I k¢"~" 1") \ I/ II / /, /" ~ I ~( ( /.,r.,...:',,",.r, iiiitl .k \ \ \ r11~II%% ~-. I l l , ~ h ...... , ~',,% y ",1 I \ m
.,-~¢.., I I ,," . . o " . . " ,," ,:
•---.-.:
.~,
10 m
i i i,,'..,"
" "" "
: ,.
',,.,,,--t ~. ,.' . . . .
f
• "..-;.-.'.,:....... ¢
%'*
2222"2.''
?
•
I
~.n:~
~ I'
ooo
.-
- . / :/ "
\
Fig. 3. Field patterns for 2 subjects over the right hemisphere during N100m to the standards and during the largest response to the early stimuli; the latencies are shown in the left upper corner of each map. Isocontour plots were constructed by projecting the sensor locations onto a spherical surface whose radius was the average distance of the pickup coils from the center of the sphere approximating the shape of the head in the measurement area. Continuous lines indicate magnetic flux out of the skull, line separation is 20 tT for S 1, and 40 iT for $4. The dots indicate the measurement locations and the arrows show the locations and orientations of the equivalent current dipoles. The x-axis of the coordinate system runs approximately along the Sylvian fissure and forms a 45-degree angle with the line connecting the right ear canal to the eye corner; the origin is 7 cm posterior to the eye corner along the x-axis.
168
In a recent study [7] rare changes in stimulus duration (from 100 to 50 ms or reverse) evoked strong responses in the human supratemporal auditory cortex. The present results show that the auditory cortex is also sensitive to changes in duration of silent intervals between the tones, both when the subject ignores the stimuli and more clearly when he pays attention to them. The neural mechanisms underlying this type of feature detection are unknown. Interestingly, cat auditory cortex contains some cells whose unit activity increases when the ISI decreases from 900 to 550 ms
[6]. The neural network generating strong responses to shortenings of the interstimulus interval can be thought to be adapted by the monotonous stimulus sequence and thereby to contain a primitive memory trace of the history of stimulation; such a trace has been earlier suggested to be formed by different physical features of the stimuli, like frequency and intensity [11]. In agreement with this interpretation, lesions of the auditory cortex in rat and monkey result in impaired discrimination of temporal ordering of stimuli, a finding that suggests deficits in the auditory sensory memory
[1, 21. This study has been supported by the Academy of Finland and by the Award for the Advancement of European Science by the K6rber Foundation (Hamburg). We acknowledge J. H~llstr6m for hardware and software design and maintenance, and O.V. Lounasmaa, M. Sams, and S.J. Williamson for comments on the manuscript. I Dewson, J.H. III, Cowey, A. and Weiskrantz, L., Disruptions of auditory sequence discrimination by unilateral and bilateral cortical ablations of superior temporal gyrus in the monkey, Exp, Neurol., 28 (1970) 529 548. 2 Diamond, I.T. and Neff, W.D., Ablation of temporal cortex and discrimination of auditory patterns, J. Neurophysiol., 20 (1957) 300 315. 3 Ford, J.M. and Hillyard, S.A., Event-related potentials (ERPs) to interruptions of a steady rhythm, Psychophysiology, 18 (1981 ) 322 330. 4 Hari, R., The neuromagnetic technique in the study of the human auditory cortex. In: F. Grandori, M. Hoke and G.-L. Romani (Eds), Advances in Audiology, Vol. 6, in press. 5 Hari, R, Pelizzone, M., M~ikel/i, J.P., Hfillstr6m, J., Leinonen, L. and Lounasmaa, O.V., Neuromagnetic responses of the human auditory cortex to on- and off-sets of noise bursts, Audiology, 26 (1987) 31 43. 6 Hocherman, S. and Gilat, E., Dependence of auditory cortex evoked unit activity on interstimulus interval m the cat, J. Neurphysiol., 45 (1987) 987 997. 7 Kaukoranta, E., Sams, M., HaN, R., H/imfilfiinen, M. and N/ifitfinen, R., Reactions of the human auditory cortex to changes in tone duration, submitted. Klinke, R., Fruhstorfer, H. and Finkenzeller, P., Evoked responses as a function of external and stored information, Electroencephalogr. Clin. Neurophysiol., 25 (1968) 119 122. t~ Knuutila, J., Ahlfors, S., Ahonen, A., H~illstr6m, J., Kajola, M., Lounasmaa, O.V., Vilkman, V. and Tesche, C., Large-area low noise seven-channel de SQUID magnetometer for brain research, Rev. Sci. Instr., 58 (1987)2145 2156. 10 Loveless, N.E., Potentials evoked by temporal deviance, Biol. Psychol., 22 (1986) 149 167. II N/i/it~inen, R. and Picton, T., The NI wave of the human electric and magnetic response to sound. A review and an analysis of the component structure, Psychophysiology, 24 (1987) 375 425.
Neuroscience Letters, 99 (1989) 164 168 Elsevier ScientificPublishers Ireland Ltd.
NSL 06009
Neuromagnetic responses of human auditory cortex to interruptions in a steady rhythm R. Hari, S.L. Joutsiniemi, M. HS.mfilfiinen and V. Vilkman Low Temperature Laboratory, Helsinki University ~[' Technology, Espoo (Finland) (Received 17 October 1988; Revised version received 19 December 1988; Accepted 21 December 1988) Key wordsv Auditory cortex: Rhythm; Magnetoencephalography; Evoked response; Man We have recorded, with a 7-channel SQUID gradiometer, evoked magnetic responses of 6 healthy humans to interruptions of a steady rhythm of 50 ms 'standard' tone bursts repeated once every 610 ms. Ten percent of the tones occurred 'too early', 410 ms after the precedingstimulus. The response to standards peaked, on average, at 90 ms and that to the early tones at 148 ms. Field patterns were dipolar during both responses and the equivalent sources agreed with activation of the supratemporal auditory cortex, at slightly different locations. The dipole moments were more than twice as strong for the early tones as tbr the standards. The results emphasize the importance of temporal stimulation patterns in activating the human auditory cortex,
Perception of rhythmical passages is essential for proper i n t e r p r e t a t i o n o f speech, music a n d other acoustic messages. A l t h o u g h cortical m e c h a n i s m s are certainly involved in coding of complex acoustic patterns, no d a t a exist a b o u t the reactions of the h u m a n a u d i t o r y cortex to deviations in timing of s o u n d stimuli. The cerebral magnetic fields are t h o u g h t to be m a i n l y generated by neural currents in the fissural cortex a n d the technique is, therefore, well suited for studies o f cortical areas e m b e d d e d in the Sylvian fissure. Earlier recordings have revealed how the evoked activity of the s u p r a t e m p o r a l a u d i t o r y cortex changes in time a n d location a n d how some characteristics of acoustic stimuli are represented in cortical f u n c t i o n a l feature maps (for a review, see ref. 4). Rare deviations of physical stimulus features, like frequency or intensity, in an otherwise m o n o t o n o u s sequence of ' s t a n d a r d ' tones, have been f o u n d to elicit an extra magnetic response, ' m i s m a t c h field' at the a u d i t o r y cortex [4, 11]. In the present work we show that i n t e r r u p t i o n s in a steady r h y t h m of identical stimuli m a y also change the activity of the a u d i t o r y cortex. Seven healthy adults (3 females, 4 males) were studied in our magnetically shielded room. D u r i n g the experiment the subject was lying o n a bed with his head supported by a v a c u u m cast. The s t i m u l a t i o n resembled that used in earlier electric evoked Correspondence: R. Hari, Low Temperature Laboratory, Helsinki University of Technology, SF-02150 Espoo, Finland. 0304-3940/89/$ 03.50 ':~71989 ElsevierScientificPublishers Ireland Ltd,
165
potential studies to too-early presentation of somatosensory, auditory and visual stimuli [3, 8, 10]. Sinusoidal 1 kHz tone bursts (rectangular envelope, duration 50 ms, intensity 78 dB SPL) were presented to the left ear through a plastic tube and earpiece. The steady rhythm of 'standard' tones having an interstimulus interval (ISI) of 610 ms was randomly interrupted by an early tone 410 ms after the previous standard. Both stimuli were physically identical and the ISI following the early tone was always 610 ms. Early tones consisted 10% of all stimuli; the number of standards after each early tone varied randomly with an even distribution between 3 and 15. Subjects either ignored all stimuli by concentrating on reading a novel or paid attention to the tones by counting the number of early stimuli. The magnetic field over the right hemisphere was measured with a 7-sensor lstorder gradiometer S Q U I D (Superconducting QUantum Interference Device) system of high sensitivity (5 - 6 t T / x / ~ intrinsic noise level) [9]. The pickup coils, separated by 36.5 mm, form a hexagonal array on a spherical surface (radius 125 mm), 18-20 mm above the scalp. The bandpass-filtered signals (3 dB points at 0.05 and 250 Hz, roll-off for the highpass filter 35 dB/decade and more than 80 dB/decade for the low-pass filter) were digitized at 1200 or 1500 Hz, and 600-900 artefact-free responses for standards and 70-100 for deviants were averaged on-line. The responses to the first standard after each deviant were not included in the average. In all experiments, the vertical electrooculogram was recorded and magnetic responses coinciding with blinks or eye movements were discarded from the average. S.E.M.s of the averaged responses were used as estimates for the total experimental noise, needed for confidence limit calculations. Typical values were 6-10 ffF for standards and 15-30 IT for the early tones. The signal channel with maximal amplitudes was chosen for further analysis in all subjects. The amplitudes were measured with respect to a 40 ms prestimulus baseline. Field maps were obtained from four subjects (S1, $3, $4, and $7) by measuring responses at 35-49 sensor locations over the right hemisphere. Equivalent current dipoles were found using a spherically symmetric conductor model with a center of symmetry estimated from the shape of the head in the measurement area; in this case the local radius of curvature was 12 cm. The fitting algorithm used to compute the equivalent dipoles took into account the exact sensor locations and orientations, measured with respect to the head [9], so that the effect of source and volume currents could be taken properly into account. Fig. 1 shows responses of two subjects over the right hemisphere in the attend condition. The responses of S1 to standards peak at 75 ms, and resemble the Nl00mdeflections reported earlier to different types of auditory stimuli [4, 5]. The responses to the early stimuli are clearly higher in amplitude than those to standards (360 vs 150 fF) and peak at 100-105 ms. Similar responses were obtained during repeated sessions. The spatial distribution of responses does not clearly differ in this subject for standard and early stimuli. In $2, the responses to the early stimuli were more complex in waveform containing two peaks 120-145 ms after stimulus onset. Polarity reversal between channels 1 and 4, for example, suggests that the measurement has
166
Subject
@
Subiect 2
/
!9
L in --
n 2 0 0 nl'~
l]
Fig. I. Responses to changes in the steady rhythm for two subjects; the approximate measurement locations over the righl hemisphere are shown in the inserts. The recording of S1 was repeated 4 times at the same location averaging 19 39 responses to the early tones (thin lines) and 150 300 to standards (thick lines). For $2 the corresponding numbers are 53 and 500. Passband 0.05 45 Hz.
been made approximately over the source. Further, the different distributions of the first and second peak indicate activation of more than one source. In all subjects the response waveforms are clearly different for standards and early stimuli (Fig. 2). During the ignore condition the mean (_+ S.E.M.) latency was 90 (+_ 5) ms for N l 0 0 m to standards and 148 (_+ 18) ms for the response to the early tones. In 5 out of 7 subjects the response to the early stimulus was larger during the attend than ignore condition, in two other subjects the amplitudes were the same within experimental noise. During attention the mean amplitude of all subjects was SUbleCt ]
Subject 2
Subject 3
Subject 4
Subject 5
Subiect 6
qubject 7
[2,! /*,',,
,~,
,..:.,.,:.,i,,
L,
f] z
F
,,,,,
; ~ ,'v,
I 200 tT
L
,'"i,
< F - - 1
~1
I - - I
F - - I
~
l
r ~ l
200 ms
Fig. 2. Responses of all subjects in ignore and attend conditions. The continuous lines indicate responses to standards and the dotted lines those to the early stimuli, The recordings are from the right frontotemporal area, over the anterior field extremum.
167
48% larger than during the ignore condition (P<0.05, t-test). In control recordings much stronger responses were evoked by decreases than increases of the interstimulus interval. Fig. 3 shows isocontour field maps for two subjects during N100m to standards and the response to early stimuli, both in the attend condition; maps are shown at time moments with the best goodness-of-fit values for the dipole model. The locations of the equivalent dipoles agree with activation of the supratemporal auditory cortex; the dipole model explained 85 97% of the field variance in the four subjects. The 3-dimensional locations of the dipoles for the two deflections differed by 6-29 mm in the four subjects without any systematic features; the location differences were statistically significant at the 95% confidence level in 3 subjects. The dipole moments were, on average, 120% stronger for the early stimuli (mean 27 nA.m) than for the standards (mean 12 nA.m).
Subject 1 110 ms .:.
Subject 4
fY
k
..,-
. "\",.
I..,-
-,-~
~y
,,,f
....... - ,-', . . . . .
a<
;
l
/,
.:~r ,b' t,,-,'/"
Z
t
"'"
• ........
[.,-,
.°
X
". I
......
.
...';
o,•"
145ms 9 "
\'-
• "',,
,,.......... i~
. . i I I
<
/ '
,
I ;
,1
°"
. %
i
.
.
........... •
i I
.
i i L,gV//,~
°
0 r I o
,,,'.,//
o-o°°
i]ll.VJ/
.° ~" I . . o " o• •
.o"
"
"
s
•
.-o-
~
-~.
I I1~ F o-~ lellll IBa •
"
I/I:
• •
/: : : m ~ . ' . ' i ~ l . l l
'11¢ ~1 I
1/o1~
,,'#:,. -" ,: llj
•
,,
P
~
I IIIII I IIIII
,. ~ / " •
•
am-,
.--'./
; i
"-. ~ ...--'.--~
A\\-
.?:,:;12. j
..,, ". •
f'
o'°
,s
I'.--:i.~%~,'~ ,. ~ : . : ~ / I k¢"~" 1") \ I/ II / /, /" ~ I ~( ( /.,r.,...:',,",.r, iiiitl .k \ \ \ r11~II%% ~-. I l l , ~ h ...... , ~',,% y ",1 I \ m
.,-~¢.., I I ,," . . o " . . " ,," ,:
•---.-.:
.~,
10 m
i i i,,'..,"
" "" "
: ,.
',,.,,,--t ~. ,.' . . . .
f
• "..-;.-.'.,:....... ¢
%'*
2222"2.''
?
•
I
~.n:~
~ I'
ooo
.-
- . / :/ "
\
Fig. 3. Field patterns for 2 subjects over the right hemisphere during N100m to the standards and during the largest response to the early stimuli; the latencies are shown in the left upper corner of each map. Isocontour plots were constructed by projecting the sensor locations onto a spherical surface whose radius was the average distance of the pickup coils from the center of the sphere approximating the shape of the head in the measurement area. Continuous lines indicate magnetic flux out of the skull, line separation is 20 tT for S 1, and 40 iT for $4. The dots indicate the measurement locations and the arrows show the locations and orientations of the equivalent current dipoles. The x-axis of the coordinate system runs approximately along the Sylvian fissure and forms a 45-degree angle with the line connecting the right ear canal to the eye corner; the origin is 7 cm posterior to the eye corner along the x-axis.
168
In a recent study [7] rare changes in stimulus duration (from 100 to 50 ms or reverse) evoked strong responses in the human supratemporal auditory cortex. The present results show that the auditory cortex is also sensitive to changes in duration of silent intervals between the tones, both when the subject ignores the stimuli and more clearly when he pays attention to them. The neural mechanisms underlying this type of feature detection are unknown. Interestingly, cat auditory cortex contains some cells whose unit activity increases when the ISI decreases from 900 to 550 ms
[6]. The neural network generating strong responses to shortenings of the interstimulus interval can be thought to be adapted by the monotonous stimulus sequence and thereby to contain a primitive memory trace of the history of stimulation; such a trace has been earlier suggested to be formed by different physical features of the stimuli, like frequency and intensity [11]. In agreement with this interpretation, lesions of the auditory cortex in rat and monkey result in impaired discrimination of temporal ordering of stimuli, a finding that suggests deficits in the auditory sensory memory
[1, 21. This study has been supported by the Academy of Finland and by the Award for the Advancement of European Science by the K6rber Foundation (Hamburg). We acknowledge J. H~llstr6m for hardware and software design and maintenance, and O.V. Lounasmaa, M. Sams, and S.J. Williamson for comments on the manuscript. I Dewson, J.H. III, Cowey, A. and Weiskrantz, L., Disruptions of auditory sequence discrimination by unilateral and bilateral cortical ablations of superior temporal gyrus in the monkey, Exp, Neurol., 28 (1970) 529 548. 2 Diamond, I.T. and Neff, W.D., Ablation of temporal cortex and discrimination of auditory patterns, J. Neurophysiol., 20 (1957) 300 315. 3 Ford, J.M. and Hillyard, S.A., Event-related potentials (ERPs) to interruptions of a steady rhythm, Psychophysiology, 18 (1981 ) 322 330. 4 Hari, R., The neuromagnetic technique in the study of the human auditory cortex. In: F. Grandori, M. Hoke and G.-L. Romani (Eds), Advances in Audiology, Vol. 6, in press. 5 Hari, R, Pelizzone, M., M~ikel/i, J.P., Hfillstr6m, J., Leinonen, L. and Lounasmaa, O.V., Neuromagnetic responses of the human auditory cortex to on- and off-sets of noise bursts, Audiology, 26 (1987) 31 43. 6 Hocherman, S. and Gilat, E., Dependence of auditory cortex evoked unit activity on interstimulus interval m the cat, J. Neurphysiol., 45 (1987) 987 997. 7 Kaukoranta, E., Sams, M., HaN, R., H/imfilfiinen, M. and N/ifitfinen, R., Reactions of the human auditory cortex to changes in tone duration, submitted. Klinke, R., Fruhstorfer, H. and Finkenzeller, P., Evoked responses as a function of external and stored information, Electroencephalogr. Clin. Neurophysiol., 25 (1968) 119 122. t~ Knuutila, J., Ahlfors, S., Ahonen, A., H~illstr6m, J., Kajola, M., Lounasmaa, O.V., Vilkman, V. and Tesche, C., Large-area low noise seven-channel de SQUID magnetometer for brain research, Rev. Sci. Instr., 58 (1987)2145 2156. 10 Loveless, N.E., Potentials evoked by temporal deviance, Biol. Psychol., 22 (1986) 149 167. II N/i/it~inen, R. and Picton, T., The NI wave of the human electric and magnetic response to sound. A review and an analysis of the component structure, Psychophysiology, 24 (1987) 375 425.