236
Electroencephalography and Climcal Neurophysiology, 1978, 45:236--243
© Elsevier/North-Holland Scientific Publishers, Ltd.
AVERAGED EVOKED POTENTIALS AND FREQUENCY MODULATION * M. KOHN, K. LIFSHITZ and D. LITCHFIELD Rockland Research Institute, Orangeburg, N.Y. 10962 (U.S.A.)
(Accepted for publication: December 20, 1977)
Among the m a n y investigations of the auditory evoked potential (AEP) in man there are surprisingly few studies that have used frequency modulated (FM) tone stimuli. Early studies by Davis et al. (1966) and others indicated that the AEP is 'non-specific' and is not frequency sensitive. Some reports, however, suggest that the AEP is frequencyspecific to some extent. A study by Butler (1972), in which two tones were presented simultaneously, showed the amplitude of the N1--P2 c o m p o n e n t of the AEP increasing progressively as the two tones became further removed from each other in frequency. This finding was interpreted by the author as an indication of the presence of two different pools of neurons each preferentially sensitive to one of the tones. In a more recent study, using a different approach, similar results showing frequency-specific components in the AEP were obtained (Shipley and Hyson 1976). R u h m (1970, 1971) used FM modulated tones in investigating the effects of stimulus rise time on the latency and amplitude of the ~E~:~13y linearly increasing and decreasing the frequency of the stimuli at various rates he f o u n d that the latency of the N1 c o m p o n e n t and the amplitude of the N1--P: c o m p o n e n t is sensitive to the rate of frequency change. He also concluded that linearly increasing frequency stimuli evoked larger responses and longer latencies than decreasing frequency * This work was supported in part by USPHS and the N.Y.S. Department of Mental Hygiene and Grant No. MH 24908.
stimuli. In addition, he reported that the AEP is evoked by the onset of upsweep or downsweep in frequency but not by the offset frequency sweep. This latter finding was first reported by Clynes (1969). Clynes also reported that the AEP at the onset of the upsweep and downsweep can be markedly attenuated if the frequency of the tone preceding it is not constant but either linearly increasing or decreasing in frequency. This paper is a report on some aspects of the AEP to different FM stimuli. These responses are of particular interest since the energy of an FM stimulus is constant. The nature of the responses to abrupt frequency increments and decrements (referred to as 'on' and 'off' responses in this report) which exist within this stimulus modality were studied. The hypothesis we tested was that the absence of a response at the offset of an upsweep or downsweep of frequency (Clynes 1969; Ruhm 1970, 1971) as well as the attenuated onset response when preceded by increasing or decreasing frequency (Clynes 1969) was related to physiological limitations in FM perception indicated by the difficulty of subjects in perceiving a clear-cut, distinct stimulus change.
Methods
Four male subjects participated in all aspects of these experiments. They had normal hearing and used no medication. Their ages ranged from 25 to 45 years. The FM stimuli were presented binaurally through
AEP AND FM STIMULI
AKG K-240 (Philips) earphones to subjects lying in an electrically shielded and sound attenuated room. Sound was adjusted for 40 dB intensity level relative to the subject's subjective threshold. For all stimuli, unless otherwise indicated, the stimuli frequency varied from 1000 to 1500 Hz. A positive pulse stimulus started at 1000 Hz, changed to 1500 Hz for the pulse duration and returned to 1000 Hz at the end of the pulse. A negative pulse stimulus started at 1500 Hz and changed to 1000 Hz for the pulse duration returning to 1500 Hz. For ramp modulated tones the frequency was swept between 1000 and 1500 Hz. When the ramp modulated stimuli were preceded by increasing or decreasing frequency tones these tones were swept at a rate of 200 Hz/sec unless otherwise stated. Interstimulus intervals varied pseudorandomly between 3 and 4 sec. The auditory conditions prior to stimulus onset lasted at least 1 sec in each experiment. The frequency range of 1000--1500 Hz was chosen because it lies on the relatively flat portion of the psychophysiological sound intensity versus sound frequency curve, as reported in the literature (Licklider 1951), and thus the possibility of introducing sound intensity changes along with frequency changes was minimized. Frequency modulated auditory stimuli, when modulated by signals having discontinuities, i.e., abrupt changes, are suspect of being contaminated by amplitude changes. The amplitude change t h a t occurs simultaneously with the frequency change 1 can often be heard as a click. This contamination of the FM stimulus by AM effects results from the electromechanical nature of the acoustic transducer such as loudspeakers or earphones. The FM stimuli were continuously monitored at the generator o u t p u t on an
1 An abrupt frequency change results in a frequency spectrum of a large number of harmonics and subharmonics, however, the spectral energy is concentrated mainly in the fundamental frequency.
237
oscilloscope, and in addition, the o u t p u t of the earphone was detected by 3 different microphones and also displayed on an oscilloscope to see the extent of AM contamination. No visible AM contamination occurred and the subjects participating in the experim e n t reported no audible clicks. Electroencephalographic activity was recorded from two bipolar electrode pairs (C~-O1 and C~-T3) and the left ear was grounded. The electrodes were silver-silver chloride; EEG amplifiers (Tektronix AM 502) had 1 M~2 input impedance, 0.1 Hz low frequency and 100 Hz high frequency 3 dB points. The amplified signals were averaged on a Mnemotron 400C CAT computer. The CAT c o m p u t e r was used to determine reaction times as well. The onset of stimulus triggered the CAT sweep, which was reset by the subject pushing a hand-held b u t t o n upon stimulus perception. For each sweep reset a c o u n t was registered in the m e m o r y address at the instant of reset, thus indicating the elapsed time between stimulus onset and stimulus perception. A distribution of reaction time values was obtained since 100 such stimuli were presented. The reaction time and EEG experiments were run under identical experimental conditions but at different times. This was done in order to avoid changes in the AEP shape by contamination from components not of interest (e.g. CNV, enhanced P300, etc.).
Results
'On' and 'off' responses FM tones, as described previously, varying from 50 msec to longer than 1 sec duration were used to investigate the 'on' and 'off' response characteristic. Fig. 1 shows the AEPs obtained to these stimuli for all 4 subjects (S1, $2, $3 and $4). The AEPs were the average of 100 responses and only the C~-O~ leads are shown since results for the Cz-T3 leads were very similar. The figure shows t h a t the response shapes, response amplitudes and
238
M. K O H N ET AL.
INCF~E
I~C~_A~Z
$3
STIM,
~
I
L.J
I
I
i
•25 SEC.
;
I
l
I
•50 SEC.
1
[ 1 SECOND
>
"~
i SECOND
~,-
Fig. 1. A E P s o f all 4 s u b j e c t s t o stimuli m o d u l a t e d b y pulses o f d i f f e r e n t d u r a t i o n s and d i r e c t i o n . The increasing stimuli varied f r o m 1000 t o 1500 Hz; t h e d e c r e a s i n g stimuli varied f r o m 1500 t o 1000 Hz. T h e d i f f e r e n t durat i o n s o f t h e m o d u l a t i n g pulse wave f o r m s are i n d i c a t e d o n t h e illustration.
AEP A N D FM S T I M U L I
239
response latencies were similar for the onset of the stimulus, regardless of whether or n o t the onset was a frequency increase or decrease. They had the general characteristics of the usual vertex potentials. 'Off' responses are noticeable in some subjects at the offset of the 0.25 sec stimulus and they are clearly distinguishable at the end of the 0.5 sec stimulus. The ' o f f ' reponses evoked by the offset of the stimulus were smaller in amplitude, although similar in shape to the 'on' responses. Fig. 2 shows the AEP of one subject ($1), the responses for the other 3 subjects were similar, to FM m o d u l a t e d tones varying repetitively between 1000 and 1500 Hz and having a d u t y cycle of 0.5. For a repetitive wave shape having only two states the d u t y cycle is defined as the ratio of time the wave shape is in one state to the total wave shape period. For example, if the two states have equal duration the d u t y cycle is 0.5. As can be seen from the figure the 'on' and 'off' response characteristics were very similar indicating that the direction of the frequency change, at least for frequencies used in this experiment, was not a factor in determining the response characteristics. Response to ramp modulated stimuli Fig. 3 shows typical responses (AEP), reaction time histograms (RT) and modulating stimulus wave shapes (STIM) used in investigating the AEP behavior to ramp modulation. The rate of change of the frequency from To for the ramp modulation was 500 Hz/sec
Cz-T3.-
_-
C z - O l ~ 20 pv
I+
Fig. 2. A typical A E P evoked by a continuous, repetitive square wave modulated tone stimulus.
I i
,
(a)
A.E.P,
(d)
i I I
R,T,
I h [ i
STIM, -
t t i i I I
(b)
(e)
A,E.P,
R.T,
STIM.
J
2Our I
Aft,P, ~
-
(f)
(C)
~
STIM. To
To
1 SECOND
Fig. 3. AEP and reaction time ( R T ) distributions to various stimulus types. As in all illustrations the stimulus shown is the m o d u l a t i n g wave shape. All ramps after To varied at 500 Hz/sec rate. Stimulus (d) is the offset p o r t i o n of stimulus (b). Stimuli (e) and (f) were preceded by slopes of 200 Hz/sec in the direction indicated.
either in the increasing or decreasing direction as illustrated. The response in (a) to the step change is shown to indicate a 'standard' AEP response amplitude and the 'standard' relatively narrow dispersion of the RT distribution to a well perceived change in stimulus frequency. Responses (b) and (c), which were evoked by ramp modulated frequency changes that were preceded by constant frequency tones, had similar response shapes; the response to descending frequency was smaller in amplitude. Correspondingly, the RT distribution for the descending frequency stimulus had a larger standard deviation, i.e., it was more widely dispersed. There was no visible AEP evoked by the termination of the increasing ramp stimulus and the associated
240
M. K O H N ET AL.
RT distribution had very wide dispersion as shown in (d). F o r the responses in (e) and (f) the ramp onset at To was preceded by freq u e n c y change at the rate of 200 Hz/sec, fr o m the opposite and from the same direction respectively as the stimulus. These freq u e n cy changes lasted for 1 sec prior to To. No visible AEPs were present and here again the RT dispersion was very wide. Table I gives the standard deviation of the R T distribution in milliseconds for all stimulus types for all subjects. The circled por t i on of the modulating wave shapes was designated as the stimulus onset and the subjects were asked to indicate p e r c ep tio n of stimulus onset by pushing a b u t t o n . The values of the standard deviations indicate that the offset of the ramp stimulus or the onset of ramp stimulus when preceded by either increasing or decreasing freq u e n c y tones resulted in an increase of the standard deviation of the RT distribution. This latter effect is shown on the amplitude o f the AEP in mo re detail in Fig. 4. This illustration shows AEPs in response to the same increasing stimulus, a positive ramp starting at To and changing in f r e q u e n c y f r om 1000 to 1500 Hz at a 500 Hz/sec rate but preceded by ramps of different slope. T he p e r c e n t label on the left-hand side of the illustration indicates that the slope of the ramp preceding To as a per cent of the stimulus slope following it. The plus sign indicates the direction of the slope prior to To; if the sign is positive it is in the same direction as the stimulus, if it TABLE I S t a n d a r d d e v i a t i o n s o f t h e R T d i s t r i b u t i o n in millis e c o n d s for all 4 s u b j e c t s ($1, $2, $3, $ 4 ) for t h e diff e r e n t s t i m u l u s t y p e s used. T h e circled p o r t i o n o f t h e wave s h a p e was d e s i g n a t e d as t h e s t i m u l u s to be identified b y t h e subjects.
A- _ J - J / -
-N
51
29,4
32.5
82.3
143.1
43,0
lO~,O
$2
23,9
79.8
IG5,2
195.4
93,6
148.5
$3
41.8
46,3
150,1
183.6
76,4
I
$4
40.1
59.8
133,3
L~9.5
50,6
I 175,9
116.7
I
I
-10% I
P
I
+5%
~
~
-
+i~
~
+20%
~
, ~ " - ' , v - " ~ ~ I I l ~ ~
I) ,
L
+
20 uv I -
~ i SECOnd)
Fig. 4. AEPs t o r a m p s t i m u l i w h e n p r e c e d e d b y r a m p s o f d i f f e r e n t slopes e x p r e s s e d as a p e r c e n t o f t h e stimulus slope. T h e positive sign i n d i c a t e s t h a t t h e direct i o n of t h e r a m p p r e c e d i n g T O is in t h e s a m e d i r e c t i o n as t h e s t i m u l u s slope; t h e negative sign i n d i c a t e s o p p o s i t e directions.
is negative it is in the opposite direction. It can be seen that the AEP was a t t e n u a t e d by either polarity ramp preceding To but those having the same direction as the stimulus, the positive ones, were more effective in attenuating the response than those having the opposite direction.
Discussion As m e n t i o n e d previously, although the m oni t ori ng of the stimulus wave shapes showed no visible AM c o n t a m i n a t i o n and the subjects did n o t hear any audible clicks the possibility t hat the AEPs were evoked partly by acoustic transient artifacts could n o t be ruled out based on these observations alone. The possibility can be ruled out, however, based on our experimental results. If the AEPs were caused by AM transients due to sudden f r e q u e n c y changes one would expect AEPs of equal magnitude to stimuli of increasing f r e q u e n c y ramps w h e t h e r t hey are preceded by decreasing or const ant f r e q u e n c y tones. In fact, when an increasing f r e q u e n c y ramp stimulus is preceded by a decreasing freo
AEP AND FM STIMULI quency ramp the discontinuity, i.e., the suddenness in frequency change, is larger than when the ramp is preceded by constant frequency. If these frequency changes cause AM contamination of the stimulus, the AEP obtained from the former should at least be equal in magnitude to the AEP evoked by the latter. Fig. 3, showing AEPs to various stimulus types, indicates this is not the case. The figure shows that stimulus case {b), an increasing frequency ramp preceded by constant frequency, evoked an AEP, while stimulus case (e), the same increasing frequency ramp preceded by a decreasing frequency tone, evoked no detectable AEP. Similarly, one would find it difficult to explain the results shown for stimulus cases (c) and (d) if one assumed that the AEPs are evoked by acoustic transient artifacts. Thus, we believe t h a t the results reported are due to the frequency modulated stimuli and n o t to amplitude modulated artifacts. Our results indicate that frequency modulated tone stimuli evoked the usual vertex potentials. For sudden, step-like frequency changes the direction of the change does n o t appreciably influence either the amplitude or the time profile of the AEP. Thus, one cannot label AEPs as an 'on' or ' o f f ' type based on the direction of the frequency change. Our results show t h a t they are determined by the d u t y cycle of the stimulus, i.e., an 'on' response is evoked whenever the frequency changes from a tone of long duration to a different frequency of shorter duration. Similarly, the ' o f f ' response occurs whenever the frequency changes from a tone of short duration to a different frequency tone of longer duration. As is shown in the previous section, whenever the two different tones have equal duration and are alternately presented in a continuous, repetitive manner the AEPs evoked at each frequency change are similar. A related finding to the above using sound and absence of sound as stimuli was reported by J~irvilehto and Fruhstorfer (1973). They showed that similar responses were evoked both by onsets of sound and by the onsets of
241 pauses when these onsets were preceded by long duration pauses and sound respectively. Our results confirm and further extend the findings of Clynes (1969) and R u h m (1970, 1971) for ramp modulated tone stimuli. We replicated the reported results that ramp modulated tone stimuli when preceded by ramp modulated tones of lesser slope will attenuate and can, when the slope is sufficiently large, abolish the AEP. Another phen o m e n o n replicated was that at the offset of ramp modulated tone stimuli, the AEP is absent. Our results indicate that both of these observations are related to the physiological difficulty subjects have in sensing these stimuli. The RT experiments indicated that the dispersion of the RT distribution, as indicated by the standard deviation, was inversely proportional to the amplitude of the AEP. Since the increased dispersion of the RT values is a reflection of the subjects' difficulty in sensing stimulus onset, it is reasonable to say that AEP attenuation and absence are related to this sensory difficulty. Whether the individual evoked potentials are attenuated or the attenuation of the AEP is a 'time smearing' p h e n o m e n o n in the averaging procedure is difficult to say. The time scatter of the RT experiments indicates smearing, however, attenuation of responses may also have occurred. We also find, as did R u h m (1971), that unlike the AEPs to step function modulated tones the AEPs obtained to negative ramp modulated tones are smaller in amplitude than those evoked by positive ramps. One possible explanation of this p h e n o m e n o n may be the number of neuronal cells involved in the processing of different stimuli. We may also note the fact that a fixed physical rate of change of stimulus frequency is perceived psychophysically as a larger rate of stimulus change at low frequencies than at high frequencies. Thus, we would expect to obtain a larger AEP amplitude for positive ramp modulated stimuli starting at 1000 Hz than for negative ramp modulated stimuli starting at 1500 Hz. Indeed, when we compared AEP
242 amplitudes to positive ramps with slopes of 500 Hz/sec with AEP amplitudes to negative ramps having slopes of 750 Hz/sec the amplitudes were similar. For the step change in stimuli the 500 Hz change occurs essentially instantaneously, so t h a t the direction of change does n o t affect the psychophysical magnitude of the tone stimulus.
Summary Frequency modulated (FM) auditory stimuli result in average vertex potentials similar to the usual auditory average evoked potential (AEP). For stepwise increase or decrease in tone frequency the AEPs are similar. For FM stimuli modulated by pulses of different durations 'on' responses are evoked by the transition of the stimulus from the longer duration to the shorter duration frequency tone while ' o f f ' responses result when the frequency transition is from the shorter to the longer duration tone. Ramp m o d u l a t i o n of the stimulus frequency results in average evoked responses; the amplitude of these responses is proportional to the slope of the ramp as well as the frequency of the tone t h a t precedes the ramp. Thus, if the tone preceding the ramp is also a ramp but of smaller slope the AEP is attenuated and with sufficiently large slope the AEP can be completely extinguished. No AEPs were obtained at the offset of ramp modulated stimuli. The standard deviation (S.D.) of the reaction time (RT) distributions to stimulus onset indicate that the AEP amplitude is inversely proportional to the S.D. values. Thus, the attenuation p h e n o m e n a appeared to be related to the uncertainty of the subject as to the exact time the stimulus occurred, both of which seem to be the result of sensory difficulty to the type of stimuli used. AEPs to negative ramps were smaller than AEPs to positive ramps; this m a y be on account of the psychological inequality between the stimuli.
M. KOHN ET AL.
R~sum~ Potentiels dvoquds auditifs et modulation de frgquence Des stimuli auditifs en modulation de fr~quence (FM) provoquent des potentiels moyens au vertex similaires aux potentiels ~voqu~s auditifs moyens habituels (AEP). Pour une augmentation ou une diminution brusque de la tonalit~ du stimulus, les AEPs sont identiques. Pour des stimuli (FM) modul~s par des pulsations de dur~es diff~rentes, des r~ponses 'on' sont ~voqu~es par le passage de la tonalit~ qui dure le plus longtemps la tonalit~ de plus courte dur~e, et des r~ponses 'off', dans le cas contraire. Une variation progressive de la tonalit~ provoque aussi une r~ponse 5voqu~e m o y e n n e ; l'amplitude de cette r~ponse est proportionnelle ~ la vitesse de cette variation (pente de la 'rampe') ainsi qu'~ la tonalit~ qui la precede. Ainsi, si la tonalit~ initiale est elle-m~me une variation progressive, mais de plus faible pente, I'AEP est att~nu~ et peut m~me ~tre compl~tement absent si cette variation initiale est suffisamment forte. Aucun AEP n'est obtenu ~ l'extinction des stimuli modul~s par variation progressive. L'~tude des distributions des temps de r~action ~ l'~tablissement des stimuli montre que l'amplitude des AEPs est inversement proportionnelle ~ la valeur de l'~cart-type de la distribution. Ainsi, l'att~nuation du ph~nom~ne apparait li~e ~ l'incertitude du sujet quant au m o m e n t exact d'occurrence du stimulus, ceci ~tant sans doute dfi ~ la difficult~ que repr~sente, pour le syst~me sensoriel, le type de stimulation utilis~. Enfin, le sens de la variation a une influence sur les AEPs: les r~ponses sont moins amples pour une variation progressive n~garive que pour une variation progressive positive; ceci est probablement en rapport avec la difference de signification psychologique de ces stimuli. The assistance of Lawrence M. Lifshitz is gratefully acknowledged.
AEP AND FM STIMULI
References Butler, R.A. Frequency specificity of the auditory evoked responses to simultaneously and successively presented stimuli. Electroenceph. clin. Neurophysiol., 1972, 33: 277--282. Clynes, M. Dynamics of vertex evoked potentials: the R-M brain function. In: E. Donchin and D.B. Lindsley (Eds.), Averaged Evoked Potentials: Methods, Results, Evaluations. National Aeronautics and Space Administration (NASA SP-191), Washington, D.C., 1969: 363--374. Davis, H., Mast, T., Yoshie, N. and Zerlin, S. The slow responses of the human cortex to auditory stimuli: recovery process. Electroenceph. clin. Neurophysiol., 1966, 21: 105--113. J~/rvilehto, T. and Fruhstorfer, H. Is the sound-
243 evoked DC potential a contingent negative variation? Electroenceph. clin. Neurophysiol., 1973, Suppl. 38: 105--108. Licklider, J.C.R. Basic correlates of the auditory stimulus. In: S.S. Stevens (Ed.), Handbook of Experimental Psychology. Wiley, New York, 1951: 985--1039. Ruhm, H.B. Role of frequency change and the acoustically evoked response. J. Auditory Res., 1970, 10: 29--34. Ruhm, H.B. Directional sensitivity and taterality of electroencephalic response evoked by acoustic sweep frequencies. J. Auditory Res., 1971, 11: 9-16. Shipley, T. and Hyson, M. On auditory and pitch specific evoked cortical potentials. J. acoust. Soc. Amer., 1976, 59: 1519--1520.
Electroencephalography and Climcal Neurophysiology, 1978, 45:236--243
© Elsevier/North-Holland Scientific Publishers, Ltd.
AVERAGED EVOKED POTENTIALS AND FREQUENCY MODULATION * M. KOHN, K. LIFSHITZ and D. LITCHFIELD Rockland Research Institute, Orangeburg, N.Y. 10962 (U.S.A.)
(Accepted for publication: December 20, 1977)
Among the m a n y investigations of the auditory evoked potential (AEP) in man there are surprisingly few studies that have used frequency modulated (FM) tone stimuli. Early studies by Davis et al. (1966) and others indicated that the AEP is 'non-specific' and is not frequency sensitive. Some reports, however, suggest that the AEP is frequencyspecific to some extent. A study by Butler (1972), in which two tones were presented simultaneously, showed the amplitude of the N1--P2 c o m p o n e n t of the AEP increasing progressively as the two tones became further removed from each other in frequency. This finding was interpreted by the author as an indication of the presence of two different pools of neurons each preferentially sensitive to one of the tones. In a more recent study, using a different approach, similar results showing frequency-specific components in the AEP were obtained (Shipley and Hyson 1976). R u h m (1970, 1971) used FM modulated tones in investigating the effects of stimulus rise time on the latency and amplitude of the ~E~:~13y linearly increasing and decreasing the frequency of the stimuli at various rates he f o u n d that the latency of the N1 c o m p o n e n t and the amplitude of the N1--P: c o m p o n e n t is sensitive to the rate of frequency change. He also concluded that linearly increasing frequency stimuli evoked larger responses and longer latencies than decreasing frequency * This work was supported in part by USPHS and the N.Y.S. Department of Mental Hygiene and Grant No. MH 24908.
stimuli. In addition, he reported that the AEP is evoked by the onset of upsweep or downsweep in frequency but not by the offset frequency sweep. This latter finding was first reported by Clynes (1969). Clynes also reported that the AEP at the onset of the upsweep and downsweep can be markedly attenuated if the frequency of the tone preceding it is not constant but either linearly increasing or decreasing in frequency. This paper is a report on some aspects of the AEP to different FM stimuli. These responses are of particular interest since the energy of an FM stimulus is constant. The nature of the responses to abrupt frequency increments and decrements (referred to as 'on' and 'off' responses in this report) which exist within this stimulus modality were studied. The hypothesis we tested was that the absence of a response at the offset of an upsweep or downsweep of frequency (Clynes 1969; Ruhm 1970, 1971) as well as the attenuated onset response when preceded by increasing or decreasing frequency (Clynes 1969) was related to physiological limitations in FM perception indicated by the difficulty of subjects in perceiving a clear-cut, distinct stimulus change.
Methods
Four male subjects participated in all aspects of these experiments. They had normal hearing and used no medication. Their ages ranged from 25 to 45 years. The FM stimuli were presented binaurally through
AEP AND FM STIMULI
AKG K-240 (Philips) earphones to subjects lying in an electrically shielded and sound attenuated room. Sound was adjusted for 40 dB intensity level relative to the subject's subjective threshold. For all stimuli, unless otherwise indicated, the stimuli frequency varied from 1000 to 1500 Hz. A positive pulse stimulus started at 1000 Hz, changed to 1500 Hz for the pulse duration and returned to 1000 Hz at the end of the pulse. A negative pulse stimulus started at 1500 Hz and changed to 1000 Hz for the pulse duration returning to 1500 Hz. For ramp modulated tones the frequency was swept between 1000 and 1500 Hz. When the ramp modulated stimuli were preceded by increasing or decreasing frequency tones these tones were swept at a rate of 200 Hz/sec unless otherwise stated. Interstimulus intervals varied pseudorandomly between 3 and 4 sec. The auditory conditions prior to stimulus onset lasted at least 1 sec in each experiment. The frequency range of 1000--1500 Hz was chosen because it lies on the relatively flat portion of the psychophysiological sound intensity versus sound frequency curve, as reported in the literature (Licklider 1951), and thus the possibility of introducing sound intensity changes along with frequency changes was minimized. Frequency modulated auditory stimuli, when modulated by signals having discontinuities, i.e., abrupt changes, are suspect of being contaminated by amplitude changes. The amplitude change t h a t occurs simultaneously with the frequency change 1 can often be heard as a click. This contamination of the FM stimulus by AM effects results from the electromechanical nature of the acoustic transducer such as loudspeakers or earphones. The FM stimuli were continuously monitored at the generator o u t p u t on an
1 An abrupt frequency change results in a frequency spectrum of a large number of harmonics and subharmonics, however, the spectral energy is concentrated mainly in the fundamental frequency.
237
oscilloscope, and in addition, the o u t p u t of the earphone was detected by 3 different microphones and also displayed on an oscilloscope to see the extent of AM contamination. No visible AM contamination occurred and the subjects participating in the experim e n t reported no audible clicks. Electroencephalographic activity was recorded from two bipolar electrode pairs (C~-O1 and C~-T3) and the left ear was grounded. The electrodes were silver-silver chloride; EEG amplifiers (Tektronix AM 502) had 1 M~2 input impedance, 0.1 Hz low frequency and 100 Hz high frequency 3 dB points. The amplified signals were averaged on a Mnemotron 400C CAT computer. The CAT c o m p u t e r was used to determine reaction times as well. The onset of stimulus triggered the CAT sweep, which was reset by the subject pushing a hand-held b u t t o n upon stimulus perception. For each sweep reset a c o u n t was registered in the m e m o r y address at the instant of reset, thus indicating the elapsed time between stimulus onset and stimulus perception. A distribution of reaction time values was obtained since 100 such stimuli were presented. The reaction time and EEG experiments were run under identical experimental conditions but at different times. This was done in order to avoid changes in the AEP shape by contamination from components not of interest (e.g. CNV, enhanced P300, etc.).
Results
'On' and 'off' responses FM tones, as described previously, varying from 50 msec to longer than 1 sec duration were used to investigate the 'on' and 'off' response characteristic. Fig. 1 shows the AEPs obtained to these stimuli for all 4 subjects (S1, $2, $3 and $4). The AEPs were the average of 100 responses and only the C~-O~ leads are shown since results for the Cz-T3 leads were very similar. The figure shows t h a t the response shapes, response amplitudes and
238
M. K O H N ET AL.
INCF~E
I~C~_A~Z
$3
STIM,
~
I
L.J
I
I
i
•25 SEC.
;
I
l
I
•50 SEC.
1
[ 1 SECOND
>
"~
i SECOND
~,-
Fig. 1. A E P s o f all 4 s u b j e c t s t o stimuli m o d u l a t e d b y pulses o f d i f f e r e n t d u r a t i o n s and d i r e c t i o n . The increasing stimuli varied f r o m 1000 t o 1500 Hz; t h e d e c r e a s i n g stimuli varied f r o m 1500 t o 1000 Hz. T h e d i f f e r e n t durat i o n s o f t h e m o d u l a t i n g pulse wave f o r m s are i n d i c a t e d o n t h e illustration.
AEP A N D FM S T I M U L I
239
response latencies were similar for the onset of the stimulus, regardless of whether or n o t the onset was a frequency increase or decrease. They had the general characteristics of the usual vertex potentials. 'Off' responses are noticeable in some subjects at the offset of the 0.25 sec stimulus and they are clearly distinguishable at the end of the 0.5 sec stimulus. The ' o f f ' reponses evoked by the offset of the stimulus were smaller in amplitude, although similar in shape to the 'on' responses. Fig. 2 shows the AEP of one subject ($1), the responses for the other 3 subjects were similar, to FM m o d u l a t e d tones varying repetitively between 1000 and 1500 Hz and having a d u t y cycle of 0.5. For a repetitive wave shape having only two states the d u t y cycle is defined as the ratio of time the wave shape is in one state to the total wave shape period. For example, if the two states have equal duration the d u t y cycle is 0.5. As can be seen from the figure the 'on' and 'off' response characteristics were very similar indicating that the direction of the frequency change, at least for frequencies used in this experiment, was not a factor in determining the response characteristics. Response to ramp modulated stimuli Fig. 3 shows typical responses (AEP), reaction time histograms (RT) and modulating stimulus wave shapes (STIM) used in investigating the AEP behavior to ramp modulation. The rate of change of the frequency from To for the ramp modulation was 500 Hz/sec
Cz-T3.-
_-
C z - O l ~ 20 pv
I+
Fig. 2. A typical A E P evoked by a continuous, repetitive square wave modulated tone stimulus.
I i
,
(a)
A.E.P,
(d)
i I I
R,T,
I h [ i
STIM, -
t t i i I I
(b)
(e)
A,E.P,
R.T,
STIM.
J
2Our I
Aft,P, ~
-
(f)
(C)
~
STIM. To
To
1 SECOND
Fig. 3. AEP and reaction time ( R T ) distributions to various stimulus types. As in all illustrations the stimulus shown is the m o d u l a t i n g wave shape. All ramps after To varied at 500 Hz/sec rate. Stimulus (d) is the offset p o r t i o n of stimulus (b). Stimuli (e) and (f) were preceded by slopes of 200 Hz/sec in the direction indicated.
either in the increasing or decreasing direction as illustrated. The response in (a) to the step change is shown to indicate a 'standard' AEP response amplitude and the 'standard' relatively narrow dispersion of the RT distribution to a well perceived change in stimulus frequency. Responses (b) and (c), which were evoked by ramp modulated frequency changes that were preceded by constant frequency tones, had similar response shapes; the response to descending frequency was smaller in amplitude. Correspondingly, the RT distribution for the descending frequency stimulus had a larger standard deviation, i.e., it was more widely dispersed. There was no visible AEP evoked by the termination of the increasing ramp stimulus and the associated
240
M. K O H N ET AL.
RT distribution had very wide dispersion as shown in (d). F o r the responses in (e) and (f) the ramp onset at To was preceded by freq u e n c y change at the rate of 200 Hz/sec, fr o m the opposite and from the same direction respectively as the stimulus. These freq u e n cy changes lasted for 1 sec prior to To. No visible AEPs were present and here again the RT dispersion was very wide. Table I gives the standard deviation of the R T distribution in milliseconds for all stimulus types for all subjects. The circled por t i on of the modulating wave shapes was designated as the stimulus onset and the subjects were asked to indicate p e r c ep tio n of stimulus onset by pushing a b u t t o n . The values of the standard deviations indicate that the offset of the ramp stimulus or the onset of ramp stimulus when preceded by either increasing or decreasing freq u e n c y tones resulted in an increase of the standard deviation of the RT distribution. This latter effect is shown on the amplitude o f the AEP in mo re detail in Fig. 4. This illustration shows AEPs in response to the same increasing stimulus, a positive ramp starting at To and changing in f r e q u e n c y f r om 1000 to 1500 Hz at a 500 Hz/sec rate but preceded by ramps of different slope. T he p e r c e n t label on the left-hand side of the illustration indicates that the slope of the ramp preceding To as a per cent of the stimulus slope following it. The plus sign indicates the direction of the slope prior to To; if the sign is positive it is in the same direction as the stimulus, if it TABLE I S t a n d a r d d e v i a t i o n s o f t h e R T d i s t r i b u t i o n in millis e c o n d s for all 4 s u b j e c t s ($1, $2, $3, $ 4 ) for t h e diff e r e n t s t i m u l u s t y p e s used. T h e circled p o r t i o n o f t h e wave s h a p e was d e s i g n a t e d as t h e s t i m u l u s to be identified b y t h e subjects.
A- _ J - J / -
-N
51
29,4
32.5
82.3
143.1
43,0
lO~,O
$2
23,9
79.8
IG5,2
195.4
93,6
148.5
$3
41.8
46,3
150,1
183.6
76,4
I
$4
40.1
59.8
133,3
L~9.5
50,6
I 175,9
116.7
I
I
-10% I
P
I
+5%
~
~
-
+i~
~
+20%
~
, ~ " - ' , v - " ~ ~ I I l ~ ~
I) ,
L
+
20 uv I -
~ i SECOnd)
Fig. 4. AEPs t o r a m p s t i m u l i w h e n p r e c e d e d b y r a m p s o f d i f f e r e n t slopes e x p r e s s e d as a p e r c e n t o f t h e stimulus slope. T h e positive sign i n d i c a t e s t h a t t h e direct i o n of t h e r a m p p r e c e d i n g T O is in t h e s a m e d i r e c t i o n as t h e s t i m u l u s slope; t h e negative sign i n d i c a t e s o p p o s i t e directions.
is negative it is in the opposite direction. It can be seen that the AEP was a t t e n u a t e d by either polarity ramp preceding To but those having the same direction as the stimulus, the positive ones, were more effective in attenuating the response than those having the opposite direction.
Discussion As m e n t i o n e d previously, although the m oni t ori ng of the stimulus wave shapes showed no visible AM c o n t a m i n a t i o n and the subjects did n o t hear any audible clicks the possibility t hat the AEPs were evoked partly by acoustic transient artifacts could n o t be ruled out based on these observations alone. The possibility can be ruled out, however, based on our experimental results. If the AEPs were caused by AM transients due to sudden f r e q u e n c y changes one would expect AEPs of equal magnitude to stimuli of increasing f r e q u e n c y ramps w h e t h e r t hey are preceded by decreasing or const ant f r e q u e n c y tones. In fact, when an increasing f r e q u e n c y ramp stimulus is preceded by a decreasing freo
AEP AND FM STIMULI quency ramp the discontinuity, i.e., the suddenness in frequency change, is larger than when the ramp is preceded by constant frequency. If these frequency changes cause AM contamination of the stimulus, the AEP obtained from the former should at least be equal in magnitude to the AEP evoked by the latter. Fig. 3, showing AEPs to various stimulus types, indicates this is not the case. The figure shows that stimulus case {b), an increasing frequency ramp preceded by constant frequency, evoked an AEP, while stimulus case (e), the same increasing frequency ramp preceded by a decreasing frequency tone, evoked no detectable AEP. Similarly, one would find it difficult to explain the results shown for stimulus cases (c) and (d) if one assumed that the AEPs are evoked by acoustic transient artifacts. Thus, we believe t h a t the results reported are due to the frequency modulated stimuli and n o t to amplitude modulated artifacts. Our results indicate that frequency modulated tone stimuli evoked the usual vertex potentials. For sudden, step-like frequency changes the direction of the change does n o t appreciably influence either the amplitude or the time profile of the AEP. Thus, one cannot label AEPs as an 'on' or ' o f f ' type based on the direction of the frequency change. Our results show t h a t they are determined by the d u t y cycle of the stimulus, i.e., an 'on' response is evoked whenever the frequency changes from a tone of long duration to a different frequency of shorter duration. Similarly, the ' o f f ' response occurs whenever the frequency changes from a tone of short duration to a different frequency tone of longer duration. As is shown in the previous section, whenever the two different tones have equal duration and are alternately presented in a continuous, repetitive manner the AEPs evoked at each frequency change are similar. A related finding to the above using sound and absence of sound as stimuli was reported by J~irvilehto and Fruhstorfer (1973). They showed that similar responses were evoked both by onsets of sound and by the onsets of
241 pauses when these onsets were preceded by long duration pauses and sound respectively. Our results confirm and further extend the findings of Clynes (1969) and R u h m (1970, 1971) for ramp modulated tone stimuli. We replicated the reported results that ramp modulated tone stimuli when preceded by ramp modulated tones of lesser slope will attenuate and can, when the slope is sufficiently large, abolish the AEP. Another phen o m e n o n replicated was that at the offset of ramp modulated tone stimuli, the AEP is absent. Our results indicate that both of these observations are related to the physiological difficulty subjects have in sensing these stimuli. The RT experiments indicated that the dispersion of the RT distribution, as indicated by the standard deviation, was inversely proportional to the amplitude of the AEP. Since the increased dispersion of the RT values is a reflection of the subjects' difficulty in sensing stimulus onset, it is reasonable to say that AEP attenuation and absence are related to this sensory difficulty. Whether the individual evoked potentials are attenuated or the attenuation of the AEP is a 'time smearing' p h e n o m e n o n in the averaging procedure is difficult to say. The time scatter of the RT experiments indicates smearing, however, attenuation of responses may also have occurred. We also find, as did R u h m (1971), that unlike the AEPs to step function modulated tones the AEPs obtained to negative ramp modulated tones are smaller in amplitude than those evoked by positive ramps. One possible explanation of this p h e n o m e n o n may be the number of neuronal cells involved in the processing of different stimuli. We may also note the fact that a fixed physical rate of change of stimulus frequency is perceived psychophysically as a larger rate of stimulus change at low frequencies than at high frequencies. Thus, we would expect to obtain a larger AEP amplitude for positive ramp modulated stimuli starting at 1000 Hz than for negative ramp modulated stimuli starting at 1500 Hz. Indeed, when we compared AEP
242 amplitudes to positive ramps with slopes of 500 Hz/sec with AEP amplitudes to negative ramps having slopes of 750 Hz/sec the amplitudes were similar. For the step change in stimuli the 500 Hz change occurs essentially instantaneously, so t h a t the direction of change does n o t affect the psychophysical magnitude of the tone stimulus.
Summary Frequency modulated (FM) auditory stimuli result in average vertex potentials similar to the usual auditory average evoked potential (AEP). For stepwise increase or decrease in tone frequency the AEPs are similar. For FM stimuli modulated by pulses of different durations 'on' responses are evoked by the transition of the stimulus from the longer duration to the shorter duration frequency tone while ' o f f ' responses result when the frequency transition is from the shorter to the longer duration tone. Ramp m o d u l a t i o n of the stimulus frequency results in average evoked responses; the amplitude of these responses is proportional to the slope of the ramp as well as the frequency of the tone t h a t precedes the ramp. Thus, if the tone preceding the ramp is also a ramp but of smaller slope the AEP is attenuated and with sufficiently large slope the AEP can be completely extinguished. No AEPs were obtained at the offset of ramp modulated stimuli. The standard deviation (S.D.) of the reaction time (RT) distributions to stimulus onset indicate that the AEP amplitude is inversely proportional to the S.D. values. Thus, the attenuation p h e n o m e n a appeared to be related to the uncertainty of the subject as to the exact time the stimulus occurred, both of which seem to be the result of sensory difficulty to the type of stimuli used. AEPs to negative ramps were smaller than AEPs to positive ramps; this m a y be on account of the psychological inequality between the stimuli.
M. KOHN ET AL.
R~sum~ Potentiels dvoquds auditifs et modulation de frgquence Des stimuli auditifs en modulation de fr~quence (FM) provoquent des potentiels moyens au vertex similaires aux potentiels ~voqu~s auditifs moyens habituels (AEP). Pour une augmentation ou une diminution brusque de la tonalit~ du stimulus, les AEPs sont identiques. Pour des stimuli (FM) modul~s par des pulsations de dur~es diff~rentes, des r~ponses 'on' sont ~voqu~es par le passage de la tonalit~ qui dure le plus longtemps la tonalit~ de plus courte dur~e, et des r~ponses 'off', dans le cas contraire. Une variation progressive de la tonalit~ provoque aussi une r~ponse 5voqu~e m o y e n n e ; l'amplitude de cette r~ponse est proportionnelle ~ la vitesse de cette variation (pente de la 'rampe') ainsi qu'~ la tonalit~ qui la precede. Ainsi, si la tonalit~ initiale est elle-m~me une variation progressive, mais de plus faible pente, I'AEP est att~nu~ et peut m~me ~tre compl~tement absent si cette variation initiale est suffisamment forte. Aucun AEP n'est obtenu ~ l'extinction des stimuli modul~s par variation progressive. L'~tude des distributions des temps de r~action ~ l'~tablissement des stimuli montre que l'amplitude des AEPs est inversement proportionnelle ~ la valeur de l'~cart-type de la distribution. Ainsi, l'att~nuation du ph~nom~ne apparait li~e ~ l'incertitude du sujet quant au m o m e n t exact d'occurrence du stimulus, ceci ~tant sans doute dfi ~ la difficult~ que repr~sente, pour le syst~me sensoriel, le type de stimulation utilis~. Enfin, le sens de la variation a une influence sur les AEPs: les r~ponses sont moins amples pour une variation progressive n~garive que pour une variation progressive positive; ceci est probablement en rapport avec la difference de signification psychologique de ces stimuli. The assistance of Lawrence M. Lifshitz is gratefully acknowledged.
AEP AND FM STIMULI
References Butler, R.A. Frequency specificity of the auditory evoked responses to simultaneously and successively presented stimuli. Electroenceph. clin. Neurophysiol., 1972, 33: 277--282. Clynes, M. Dynamics of vertex evoked potentials: the R-M brain function. In: E. Donchin and D.B. Lindsley (Eds.), Averaged Evoked Potentials: Methods, Results, Evaluations. National Aeronautics and Space Administration (NASA SP-191), Washington, D.C., 1969: 363--374. Davis, H., Mast, T., Yoshie, N. and Zerlin, S. The slow responses of the human cortex to auditory stimuli: recovery process. Electroenceph. clin. Neurophysiol., 1966, 21: 105--113. J~/rvilehto, T. and Fruhstorfer, H. Is the sound-
243 evoked DC potential a contingent negative variation? Electroenceph. clin. Neurophysiol., 1973, Suppl. 38: 105--108. Licklider, J.C.R. Basic correlates of the auditory stimulus. In: S.S. Stevens (Ed.), Handbook of Experimental Psychology. Wiley, New York, 1951: 985--1039. Ruhm, H.B. Role of frequency change and the acoustically evoked response. J. Auditory Res., 1970, 10: 29--34. Ruhm, H.B. Directional sensitivity and taterality of electroencephalic response evoked by acoustic sweep frequencies. J. Auditory Res., 1971, 11: 9-16. Shipley, T. and Hyson, M. On auditory and pitch specific evoked cortical potentials. J. acoust. Soc. Amer., 1976, 59: 1519--1520.
