Citation
Developmental differences in the temporal summation of transient and sustained auditory stimuli

Material Information

Title:
Developmental differences in the temporal summation of transient and sustained auditory stimuli
Creator:
Blumenthal, Terry D., 1954-
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Copyright Date:
1985
Language:
English

Subjects

Subjects / Keywords:
Acoustic reflex ( jstor )
Adults ( jstor )
Amplifiers ( jstor )
Infants ( jstor )
Mental stimulation ( jstor )
Neonates ( jstor )
Philosophical psychology ( jstor )
Psychology ( jstor )
Rats ( jstor )
Reflexes ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Terry D. Blumenthal. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
AEH0186 ( ltuf )
14557916 ( oclc )
0029499517 ( ALEPH )

Downloads

This item has the following downloads:


Full Text















DEVELOPMENTAL DIFFERENCES IN THE TEMPORAL SUMMATION
OF TRANSIENT AND SUSTAINED AUDITORY STIMULI






By


TERRY D. BLUMENTHAL


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA


1985





















Dedicated to my parents,

Ella Dora Blumenthal and Otto John Blumenthal, for giving me the tools I need to get the job done.
















ACKNOWLEDGEMENTS


I would like to thank the members of my doctoral committee, Drs. W. Keith Berg (Chair), Kenneth J. Gerhardt, Peter J. Lang, Patricia H. Miller, and Donald C. Teas, for their help In bringing this work into being and for their generous donations of time and ideas. It is always helpful to have interested and learned experts to provide comments and critique. I would further like to thank Dr. Berg for his unfailing friendship and guidance throughout my graduate career. His constant and energetic drive to achieve quality and precision In research and teaching have been the ideals I compare my own endeavors to, In order to improve my skills. I would also like to thank Kathy Berg for her helpful advice, and for writing such an interesting dissertation.

I would like to thank Margarete Davies, Barry Hurwitz, Kristen Bomas, and Alonso Avendano for all their help in the gathering and analysis of the data. I would further like to thank Alonso Avendano for his patience and hard work beyond the call of duty. If not for his help, I would probably still be trying to collect these data. I would also like to thank all of the undergraduates and other adult subjects, the staff of the Newborn Nursery at Shands Hospital in Gainesville, Florida, and the parents and their new bundles of joy, for their patience and cooperation.

I am grateful to the students, faculty, and staff of the Department of Psychology at the University of Florida, for making









graduate school such an intensely enjoyable experience. I am also grateful to the faculty, staff, and students of Hamilton College for showing me that there is life after graduate school. Finally, I would like to thank Dr. Berg again, for setting such a good example.
















TABLE OF CONTENTS


ACKNOWLEDGEMENTS . III ABSTRACT . vi CHAPTER

1. INTRODUCTION . 1

Transient and Sustained Systems . 1 The Startle Response . 6

11. EXPERIMENT I . 15

Method . 15 Results . 18 Discussion . 22

111. EXPERIMENT 11 . 26

Introduction . 26 Method . 26 Results . 26 Discussion . 30

IV. EXPERIMENT III . 34

Method . 34 Results . 36 Discussion . 40

V. GENERAL DISCUSSION . 43 REFERENCES . 52 BIOGRAPHICAL SKETCH . 57















Abstract of Dissertation Presented to the Graduate School
of the University of Florida In Partial Fulfillment of the
Requirements for the degree of Doctor of Philosophy DEVELOPMENTAL DIFFERENCES IN THE TEMPORAL SUMMATION
OF TRANSIENT AND SUSTAINED AUDITORY STIMULI

By

Terry D. Blumenthal

August 1985

Chairman: W. Keith Berg
Major Department: Psychology


The present study measured the startle eyeblink electromyographic (EMG) response to transient and sustained auditory stimuli in human adults and newborns, to assess immaturity of temporal summation In newborns. Stimuli were broadband noise bursts (20Hz-20KHz) with fast (<0.1 ms) rise times. In Experiments 1 and 2 adults received 95 dB(A) stimuli, in 13 stimulus conditions. One stimulus was 3 ms long, six varied In duration from 20 to 100 ms in Experiment 1 and from 30 to 55 ms in Experiment 2, and six consisted of pairs of 3 ms stimuli at onset-to- onset intervals analogous to the durations used. In

Experiment 3, newborn infants (7-36 hours old) received 100 dB(A) stimuli analogous to those of Experiments 1 and 2. Startle amplitude increased for single and paired stimuli up to 40 ms In adults and 50 ms In infants, above which single stimuli maintained this response amplitude while the second pulse In a pair was not effective. Within these summation windows, single stimuli were not significantly more









effective than analogous paired stimuli. Startle probability Increased for single stimuli up to 20 ms in adults and 35 ms in infants, above which no significant age difference was found. Single and paired stimuli were equally effective up to 45 ms for adults and 20 ms for infants. For paired stimuli, intervals of 20-45 ms were equally effective in adults, and intervals of 20-50 ms were equally effective in Infants, with the second pulse at longer Intervals having no effect, but probability for infants was lower than that for adults at all intervals tested. Response latency decreased as stimulus duration increased to 20 ms, for adults but not for Infants, and no significant effect of interpulse interval on latency was found at either age. The window for transient summation is narrower in adults (40 ms) than In infants (50 ms), and brief stimuli are more effective in adults than in infants. These data suggest that the transient system is immature In Infants and that these developmental differences are illustrated In different ways by startle amplitude, probability, and latency.















CHAPTER I
I NTROD UCT ION

Transient and Sustained Systems

The processing of sensory input may occur in different ways for transient and sustained stimuli. Transient stimuli activate onsetsensitive response systems (detection), while sustained stimuli activate response systems which are sensitive to more prolonged stimulus attributes (identification). A stimulus with both transient and sustained properties may activate both types of response system. This transient-sustained dichotomy has been found for auditory (Graham, 1979), tactile (VerrIllo, 1968), and visual (Schwartz & Loop, 1984) input, and underlying neurological mechanisms have been proposed in each modality (Gersuni, 1971; Gescheider, 1976; Hickey, 1977). Also, the mechanisms underlying this dichotomy are believed to develop at different rates, at least in the auditory (Graham, Anthony, & Zeigler, 1983) and visual systems (Hickey, 1977; Norman, Pettigrew, & Daniels, 1977).

One way in which transient and sustained systems differ Is In their temporal summation abilities, with rapid summation occurring In transient systems, and longer-lasting summation occurring in sustained systems. In the auditory system, temporal summation has been Investigated using behavioral scaling (Zwislocki, 1969), single unit responding (Gersuni, 1971), and brainstem reflex elicitation (Marsh, Hoffman, & Stitt, 1973). The present study used the acoustic startle










reflex to assess developmental differences in temporal summation in human transient and sustained systems.

Temporal summation involves the integration of sensory input over time. This is illustrated by the fact that the effectiveness of a stimulus can be increased by increasing its duration, to a point. The duration above which no further summation occurs is called the critical duration, and is the point at which the summation function (perception plotted as a function of stimulus duration) reaches asymptote. Also, a pair of stimuli can be more effective than a single stimulus, and this increases as interstimulus interval decreases. The interval below which further reductions of interstimulus interval have no greater effect can be considered a critical Interval, analogous to the critical duration. Both the critical duration and the critical interval are the points at which their respective functions of sensation plotted as a function of time reach asymptote. Temporal summation has been found in several sensory modalities, using a variety of experimental techniques.

Psychophysical investigations of visual perception show that brightness increases as stimulus duration increases to 500 ms, but not beyond (Stevens, 1966). As stimulus intensity increases, this critical duration decreases, but the slopes of the summation functions are parallel across a range of intensities. This suggests that summation occurs until some specific amount of Input has occurred, and this amount is reached more rapidly by stimuli of higher intensity. Temporal summation may asymptote due to sufficient "area under the curve", as determined by the stimulus envelope.

Schwartz and Loop (1984) have shown that visual perception may involve two partially independent systems, a chromatic and an









achromatic system. The chromatic system is sensitive to stimulus duration, showing summation functions with critical durations of 500 ins. The achromatic system is believed to be sensitive to transient stimulation only, showing no effect of duration In the range tested (43-1000 ins). These systems may be differentiated at the neuronal level, in that retinal ganglion cells with transient discharge patterns CY-cells) are more common In the periphery, where achromatic stimulation is more pronounced (Hirsch & Levanthal, 1978). Gangl ion cells with sustained discharge patterns (X-cel Is) are more common In and around the fovea, where chromatic activation is maximal. This indicates that the differential temporal summation of chromatic (sustained) and achromatic (transient) systems may be due to underlying neurological distinctions.

For intermittent visual stimuli, brightness decreases as interstimulus Interval increases beyond the critical flicker fusion threshold (Bartley, 1969). Below this threshold, temporal summation occurs to prevent the brightness from decreasing when the stimulus is not on continuously. In the visual system, the critical Interval is Illustrated by the critical flicker fusion threshold.

Temporal summation has also been shown for both taste perception (Bujas & Ostojcic, 1939, as cited in Marks, 1974), and proprioception (Brown, 1966, 1968). Brown rotated subjects, varying the duration and acceleration of the rotation. The perceived magnitude of the rotation increased as the duration of the rotation increased, at a range of accelerations. As acceleration was increased, the effect of duration on perceived magnitude of rotation reached asymptote earlier, similar to









the finding on the visual system that critical duration decreases as stimulus intensity increases.

For tactile stimuli, two patterns of temporal summation exist. If Pacinlan corpuscles are present, summation occurs with a critical duration of 500-1000 ms (Gescheider, 1976; Verrillo, 1968). This critical duration decreases as stimulus intensity Increases, just as in the visual and proprioceptive systems. If Pacinian corpuscles are not present, summation still occurs, but with a critical duration of less than 1 ms (Higashlyama & Tashiro,1983). This is analogous to the rapidly summating achromatic (transient) and slowly summating chromatic (sustained) systems found in vision. Both visual and tactile systems are distinguished by functional differences at the periphery which result in qualitatively different stimulus processing.

In the auditory system, psychophysical methods have been used extensively to Illustrate temporal summation. Loudness matching has been used to show that increasing stimulus duration results In decreased threshold, up to a critical duration of 200 ms (Scharf, 1978). Critical duration decreases as stimulus Intensity increases (Richards, 1976), but the slope of the summation function Is the same across a wide range of intensities (Small, Brandt, & Cox, 1962; Stevens & Hall, 1966), as is the case for tactile, proprIoceptive, and visual stimuli. For pairs of brief pulses, threshold increases as interpulse interval Increases, up to 200 ms (Scharf, 1978; Zwislocki, 1960). Using psychophysical methods, the critical duration and critical interval for auditory temporal summation appear to be the same.

While the use of loudness matching in assessing temporal summation has provided much information, this technique is not without problems.









Differences in subject criteria used in the matching tasks can result in variable findings. For example, when subjects are told to match on the basis of the loudness at the onset of a stimulus as opposed to the total loudness of the stimulus, differences in stimulus duration effects are found (Zwislocki & Sokolich, 1974). These problems could be avoided by using techniques which are less susceptible to cognitive influences, such as psychophysiological or neurophysiological responses.

The investigation of auditory temporal summation at the neurophysiological level has led to the identification of two types of neurons in the auditory pathway which differ in their characteristic discharge patterns to stimuli (Gersuni, 1971; Radionova, 1971). Short time constant (STC) units illustrate a transient discharge at stimulus onset (and at offset, in some cases), while long time constant (LTC) units discharge in a sustained fashion, throughout the duration of the stimulus. These units are also sensitive to different stimulus

parameters, with LTC units being sensitive to stimulus duration, but not to rise time, while the opposite is true of STC units. The two types of units show different patterns of temporal summation, with STC units having very brief critical duratlons(a few milliseconds), and LTC units having much longer critical durations. These units are found at all levels of the auditory pathway, and may be organized into

functional systems resulting in differential processing of transient and sustained stimulus attributes. In this respect, the LTC and STC systems are analogous to Pacinian and non-Pacinian systems in tactile processing, and to chromatic (X-cell) and achromatic (Y-cell) systems in visual processing. All three modalities illustrate distinctive









transient and sustained systems which possess different temporal summation properties.

Evidence for auditory temporal summation has also been found at the level of the receptor (hair cell) in monkeys (Kletsky & Stengrevics, 1972). Period histograms of receptor activity, which are correlated with positive excursions of the stimulus waveformV decay more slowly than does the stimulus, with a lag of .1 to .2 ms. This implies that some integration may be taking place at the hair cell, resulting in a prolongation of the half-wave rectified receptor potential.

Temporal summation appears to be a fundamental property of several sensory systems. This summation may vary with the stimulus envelope, until some maximal amount of summation has occurred, since critical duration decreases as stimulus intensity increases, in several sensory modalities. Also, in the tactile, visual, and auditory systems, differential summation of transient and sustained stimuli can be found, suggesting the possibility of analogous underlying mechanisms in the three systems. Investigations of such mechanisms would be aided by a response which is sensitive to small stimulus changes, and which can be found across a range of ages and species. The startle reflex is such a response.


The Startle Response

The use of the startle response as a means to assess temporal summation can be recommended for several reasons. This response is very reliable, shows considerable plasticity and sensitivity to stimulus conditions, and is mediated by a relatively simple neural circuit.









Also, the startle response is time-locked to stimulus onset, and can be measured noninvasively. A wealth of developmental and comparative data exists, illustrating the generalizability of the startle response across studies, subjects, and species.

Landis and Hunt (1939) found that the eyeblink is the first component of the startle response to appear, and that this reflex eyeblink is very reliable, resisting habituation. The reflex eyeblink is also a sensitive measure, in that it varies predictably with very small stimulus changes (Davis & Heninger, 1972).

Lesion and electrical stimulation studies in the rat have shown that the acoustic startle response is mediated by a relatively simple neural circuit (Davis, Gendelman, Tischler, & Gendelman, 1982). Auditory input ascends via the auditory nerve and the ventral cochlear nucleus to the nuclei of the lateral lemniscus, which connect to the nucleus reticularis pontis caudalis, and this is the beginning of the motor output component of the startle pathway. Startle is a brainstem reflex that can be found in decerebrate animals (Davis & Gendelman, 1977; Fox, 1979), and in anencephalic infants (W. K. Berg, unpublished data).

Startle has been measured in adult and infant animals and humans, using a variety of techniques. In animal research, the two most common measures are electromyographic (EMG) activity of the flexor muscles of the leg, and the whole-body jerk, as measured by an accelerometer in a rigidly suspended cage. In humans the eyeblink component is assessed, either by measuring the movement of the eyelid with a potentiometer, or by measuring the EMG activity of the orbicularis oculi, the muscle responsible for the eyeblink. The advantage of using the eyeblink









component of startle in the human is that it has a lower threshold than either flexor EMG or whole-body jerk, so minimally-startling stimuli can be used. The EMG measure may be preferred to lid movement, since EMG activity can be found In the absence of lid movement, whether the eyes are open or closed, making EMG both more sensitive and more versatile than lid movement.

Startle responses can be assessed by the application of two different paradigms. First, the direct effects of stimuli on the startle response can be studied, by varying the eliciting stimuli and measuring the resultant changes in the startle response. Second ' the modification of the startle response by a nonstartling stimulus (called a repulsee, occurring before the eliciting stimulus, can be measured. The first technique is aimed at a direct assessment of the startle system, while the second technique uses the startle stimulus as a probe to Investigate processing of the repulse.

The degree to which a stimulus will elicit startle is affected by several factors. Increasing stimulus rise time results in higher startle thresholds, In both rats (Fleshler, 1965) and human adults (Berg, 1973). Increasing rise time also results in smaller blinks for suprathreshold stimuli in human adults (Blumenthal & Berg, 1984) and infants (Blumenthal & Berg, 1982). Increasing stimulus intensity yields larger responses in adults (Blumenthal & Berg, 1984) and In rats (Marsh, Hoffman, & Stitt, 1973). Increasing stimulus duration up to 8 ms, but not beyond, yields larger responses in rats (Marsh, Hoffman, & Stitt, 1973), and this critical duration decreases as stimulus Intensity increases. In human adults, the critical duration Is longer than it is in rats. Berg (1973) increased stimulus duration up to 32









ins, and found a continuing increase in response amplitude. Yamada (1983) found that response amplitude increased as stimulus duration Increased to 30-50 mns, but not beyond. This illustrates a critical duration for human adults which is considerably higher than that for rats. The critical duration for infants has not previously been reported.

In rats, Increasing the interval between two brief (1 ins) pulses to 3 mns (onset-to-onset) results in larger responses, with a return to the level of a single 1 ins pulse when interpulse interval is increased to 6 ins (Marsh, Hoffman, & Stitt, 1973). The increase In responding as interpulse interval Is increased to 3 ins may be due to a reduction of neural refractoriness and more complete peripheral recovery. At interpulse intervals greater than 6 ins, the second pulse appears to have no effect, illustrating a critical interval for summation of transients of 6 ins in rats.

The startle modification technique uses the startle stimulus as a probe, and can be used to assess stimulus processing In several modalities. This method Involves the presentation of the stimulus to be assessed, the prepulse, some brief time before the startle stimulus. Depending on the stimulus parameters of the prepulse, the startle response can be inhibited or facilitated in several ways. The amount of modification Is wholly dependent upon the parameters of the prepulse, with the startle-eliciting stimulus having no effect on the modification (Stitt, Hoffman, & Marsh, 1976).

The time between onset of the prepulse and onset of the startle pulse Is called lead time, and this is an important factor in determining modification effects. At lead times of 10 ins or less in the









rat, or 30 ms or less in the human, a prepulse above detection

threshold will cause a decrease In the latency of the subsequent startle response (Hoffman & Ison, 1980). At lead times of 30-240 ms in both rats and humans, startle amplitude is reduced by the prepulse (Hoffman & Ison, 1980). The optimal lead time, where the amount of inhibition of amplitude is maximal, is about 120 ms In adults. If lead time Is equal, prepulses which remain on throughout the lead time are no more effective than those which have a duration of 20 ms (Graham & Murray, 1977). However, as the duration of transient prepulses increases, more amplitude inhibition is found. This may be because the onset and offset of the transient prepulse become more independent as they are separated in time, so their combined effect is increased
,
(Harbin & Berg, 1984).

In infants, the transient-sensitive system is believed to be immature (Graham, Strock, & Zeigler, 1981). For maximal inhibition of startle amplitude by a transient prepulse in 6-to-9-week-olds, lead time must be 225 ms, as compared with 120 ms in adults. More intense prepulses are required with these infants, and the prepulses must be




* If a brief, transient stimulus occurs at a very long lead time, it will have no effect, unless attention is directed to the prepulse (as indicated by heart rate deceleration) (Bohlin, Graham, Silverstein, & Hackley, 1981). If attention is directed towards the sensory modality in which the startle stimulus occurs, the startle response will be facilitated; if attention is directed away from the startled modality, inhibition occurs. The attentional effect can combine with the effect of a sustained prepulse, resulting in more facilitation than to either effect alone (Bohlin, Graham, Silverstein, & Hackley, 1981). The reason for this attentional effect may be that heart deceleration, or orienting, may cause an enhancement of stimulus input so that the effectiveness of a startle stimulus occurring during orienting to a prepulse is increased (Lacey & Lacey, 1974).









slightly longer (25 - 30 ms in infants, 20 ms In adults) in order to be effective (Zeigler, Strock, & Graham, 1979). Even at optimal lead times, inhibition Is less marked in infants than in adults (Graham, Strock, & Zeigler, 1981). Since startle inhibition is argued to be a function of transient system activity, these findings suggest that the temporal summation of transients Is immature in infants.

At very long lead times (2-4 sec) a prepulse which stays on throughout the lead time (sustained prepulse) can cause an increase In startle amplitude in adults and infants, but optimal lead time Is longer for infants (4 sec) than for adults (2 sec). However, the degree of facilitation Is similar at the two ages, supporting the belief that the sustained processing system is relatively mature in Infancy (Graham, Strock, & Zeigler, 1981). In rats, startle inhibition and facilitation have been Investigated at the neurochemical level and appear to be due to the activation of different receptor sites. The receptors responsible for facilitation of startle appear to mature earlier than those responsible for inhibition (Gallager, Kehne, Wakeman, & Davis, 1983). If an analogous arrangement exists in the human, developmental differences in startle modification may be expected.

Immaturity of the transient system has been proposed, based on evidence using the startle modification paradigm (Graham, Strock, Zeigler, 1981). A more direct assessment of the transient system may be possible by using the startle elicitation paradigm, Investigating stimulus effects on the startle response directly, rather than its modification by a prepulse (Dykman & Ison, 1979). Also, If transient system immaturity is believed to exist, it may be most apparent in









newborns, so this age group, rather than older infants, should be studied.

In the present study, the temporal summation of transient and sustained stimuli was assessed in human adults and newborns, using the startle elicitation technique. The stimuli used were either single stimuli, varying in duration (3-100 ms), or pairs of brief (3 ms) stimuli, with onset-to-onset intervals analogous to the duration of the single stimuli. Since peripheral recovery requires more than 3 ms (Green, 1973), the offset of very brief pulses may have a minimal effect, so the paired stimuli were treated as two transients rather than four. These stimuli were used to assess transient system activity directly.

The single stimuli longer than 3 ms included transients (at onset and offset) and a steady-state portion (between onset and offset) so that responding to these stimuli could be contributed to by both transient and sustained system activity. The contribution of the sustained system could be assessed In three ways, all of which involve the assumption that the onset of single stimuli has an effect equivalent to that of the first pulse In a pair of brief pulses, that is, transient system activity. The assessment of sustained system activity differs in how the stimulus components following onset (i.e. steady-state portion and offset) affect responding.

Stimulus offset may be assumed to have no effect on the startle response, if startle is a function of a sufficient Increase in stimulus Intensity occurring within a sufficiently brief time, as Graham (1975) states. Therefore, subtracting the response to a 3 ms pulse from that to a longer pulse (subtraction of onset transient) would yield the









effect of the steady-state portion of the stimulus. However, the influence of stimulus offset on startle has not been shown to be irrelevant, so Ignoring this possible effect may result In an

overestimation of sustained system activity (since some of the response remaining may be due to transient system activity).

A second way to infer sustained system activity is to subtract the response to a pair of transients from the response to an analogous single stimulus. This assumes that offset is as powerful a transient as onsetP and that startle is determined by sufficient stimulus change in either direction. With this method, sustained system activity will be underestimated if offset is not as effective a transient as onset (since transient system activity will be overestimated).

Since the effect of stimulus offset has not been quantified, it may be useful to consider the situation where the effect of the offset transient is greater than zero but less than that of the onset (or the 3 ms pulse). If a single stimulus is no more effective than an analogous pair of stimuli, it is assumed that the steady-state portion and offset of the single stimulus are no more effective than the second pulse in a pair. The degree to which single stimuli are more effective than paired stimuli indicates sustained system activity, though the full effect cannot be ascertained unless offset transient effects can be determined. However, if stimulus duration Increases and responding does not change, sustained system activity may have a minimal

contribution. This does not necessarily mean that sustained activity is not occurring, it just means that this activity is not contributing much to the response. When both transient system activity and sustained system activity are present, their contributions to the response may be









unequal. If the activation of the more efficient system is decreased, the activity of the other system may become more apparent The only case in which the contribution of sustained system activity would be ambiguous is that for paired stimuli which produce larger responses than single stimuli. The lower responding to single stimuli could be due to either reduced sustained system activity or to reduced transient system activity at onset.

To assess the activity of the sustained system more directly, research using stimuli with minimal transients (long rise and fall times) is needed. The effects of stimulus onset alone could be examined with fast-rising, slow-falling stimuli, so that no offset transient occurs. Stimulus offset effects could be isolated with slow-rising, fast-falling stimuli, to minimize the onset transient. These avenues of research are suggested to explore the sustained processing system more fully.















CHAPTER 11
EXPERIMENT 1





Subjects were 16 adults, 6 males and 10 females, with a mean age of 20 years (range=18 to 27 years), selected from a university undergraduate subject pool, who reported no history of hearing loss and no use of medications on the day of testing.


Stimuli

Broadband noise bursts (20Hz-2OkHz), at 95 dBCA), with a rise-fallI time of less than 0.1 ms were used. Seven of the stimuli varied in duration (3, 20, 35, 50, 65, 80, and 100 ins), and six others consisted of pairs of 3 ins pulses at intervals analogous to the durations used (20, 35, 50, 65, 80, and 100 ins, measured from onset to onset). lntertrial interval ranged from 25-35 sec, averaging 30 sec.


Apparatus

Stimuli were produced by a Grason-Stadler 455C noise generator gated through an Iconix electronic switch and a Sansul AU517 amplifier, and presented through a JBL Decade 26 loudspeaker, located approximately 1.5 m in front of the subjects. Stimulus intensity was calibrated using a General Radio 15510 sound level meter and monitored with a Hewlett-Packard 400E AC voltmeter.









Peni-orbital electromyographic (EMG) responses were collected using miniature Beckman blopotential electrodes (Ag/AgCI) filled with Synapse conducting paste. The EMVG signal was amplified by a Coulbourn Hi-Gain Bioamplifier/Coupler, with filters passing frequencies of 90250 Hz, and a gain of 63,000. The signal was then integrated with a Coulbourn Contour Following Integrator at a time constant of 80 ms (integrated EMG). The Integrated EMO was then recorded on a Beckman R411 polygraph, as well as being digitally sampled (10 bit accuracy) by a PDP-8 computer every millisecond for 250 ms after stimulus onset.




The experimenter explained the procedure and the subject was asked to read and sign an informed consent statement, and to fill out a background questionnaire. The experimenter then cleaned the area just below the left eye with a cotton swab soaked in alcohol, and two electrodes were attached, one below the center of the eye, and the other Immediately temporal to the first, as close to the orbital ridge as possible without impairing eye movement. The subject was then seated In the testing room and asked to move as little as possible, especially the eyes and head. Data were used only from trials on which an experimenter, watching the subject on closed circuit television, judged the eyes to be open. If a stimulus was presented when the eyes were closed, or during movement of the head or eyes, the trial was rejected




* The integrator time constant was determined by calibration, not by the setting on the Integrator itself. It was found that the time constant settings of two integrators of the same model varied widely, so the only reliable way to determine the actual time constant was to calibrate each separately.









by an experimenter blind to stimulus conditions. An average of 7.5 percent of the trials were rejected for each subject, and rejection was not more likely for one stimulus condition than for any other. A 13x13 Latin square was used to determine stimulus order, with the row of entry into the square being randomly determined. A session was terminated when six potentially scramble responses In each stimulus condition were obtained.


Q&td Analysis

Response amplitude was measured as the difference between response onset and peak. Response onset was judged only during the window between 20 and 100 ms after stimulus onset. Initiation of a response was judged on the basis of a monotonic increase in EMG that continued for at least 20 ms. Onset latency was then judged as the point where this increase exceeded the random variability in the 20 ms prior to this increase. The peak of the response was the first point following onset where the slope of the integrated EMG signal equalled or passed through 0, when this was followed by at least 10 ms with no further slope reversal. Response latency was measured as the time from stimulus onset to response onset.

Response probability was measured as the degree to which responses actually occurred when subject variables were optimal (i.e. eyes open, no head movement). If the integrated EIVIG signal during the response window did not deviate beyond the random noise present prior to the onset window, a failure to respond was recorded. Due to the

considerable sensitivity of the computer sampling and amplifier, the degree to which very small responses were not detected was minimal.









The response parameter most often reported by researchers In this area Is average magnitude, which is calculated from trials on which a response could have been recorded, whether a response actually occurred or not (e.g. Graham & Murray, 1977). Magnitude changes can be due to changes In either response probability or amplitude (Prokasy & Ebel, 1967),, and these two may be partially Independent measures of startle responding (Blumenthal & Berg, in press). Therefore, the present study examined response amplitude and probability separately. For each subject, response amplitude and latency were averaged across trials for each condition. ANOVAs including stimulus type (single or paired) and time (duration or interval in msec) as within-subject variables were conducted. To balance the analyses, data from the 3 ms single stimulus condition were duplicated for use in paired stimulus comparisons, as a substitute for a pair of pulses at an interval of 0 ms.



Results


Amplitude

The effect of increasing stimulus duration or interpulse interval on response amplitude differed, as illustrated by a significant stimulus type by time interaction, E (6,90)=8.88, g<.001 (see Figure 1). Increasing stimulus duration to 50 ms, but not beyond, resulted in larger responses, as shown by significant linear, E (1,15)=18.23, .R<.001, and quadratic, E (1,15)=15.37, p<.001, trends. Responses were larger at 20 than at 3 ms duration., � (15)=4.13, p<.001, and larger at 50 than at 35 ms duration, � (15)=2.64, p-<.025. No other pairwise comparisons reached significance, indicating that increasing duration











55 50



45


40 35 30 25


20


SINGLE


















PAIRED 100


3 20 35 50 65 80


DURATION/ INTERVAL (ms)

Figure 1. Adult response amplitude as a function of time for single and paired stimuli.







20

above 50 ms had no significant effect. Increasing interpulse Interval to 35 ms, but not beyond, resulted in larger responses, yielding a significant quadratic trend, E (1,15)=5.61, ja<.05, and a marginally significant cubic trend, F- (1,15)=4.46, pL<.052. Responses were larger for a pair of pulses at 20 ms interval than for a single 3 ms pulse, �

(15)=2.20, p.<.05. Increasing Interval from 3 to 35 ms resulted In larger responses, � (15)=2.85, p.<.025, but responses were not significantly larger at 35 than at 20 ms interval. At Intervals of 50 ms or more, responses were not significantly larger than for a single 3 ms pulse. Up to 35 ms, a single stimulus was not significantly more effective than a pair of transients, showing that, even with much less total energy, a second transient was as effective as the sustained portion and the offset of a single stimulus. This indicates an asymmetrical input of transient and sustained stimuli to the startle system, with transient input being more effective than sustained input, up to 35 ms. Above 35 ms, the second transient had little effect, whereas the sustained portion of the stimulus increased in

effectiveness, reaching asymptote at 50 ms. Probability

For response probability, the effect of increasing stimulus duration or interpulse interval also differed, as illustrated by a significant stimulus type by time Interaction, E (6,90)=5.34, p-<.001 (see Figure 2). A linear effect of duration appeared, E (1,15)=13.29, .R<.01, and this was due mainly to increasing duration from 3 to 20 ms, � (15)=2.22, R<.05. In fact, if the 3 ms data were excluded from the analysis, no significant duration effect was found, F (1,15)=2.12. For







































I


100


I 1 1
20 35 50


DURATION / INTERVAL (ms)

Figure 2. Adult response probability as a function of time for single and paired stimuli.


I


100


SINGLE








PAIRED


90 80 70 60 50


3


80









paired stimuli, a quadratic trend appeared, E (1,15)=6.16, P-<.025. Increasing interpulse interval to 50 ms had no further effect. At longer Intervals, the contribution of the second pulse disappeared, with probability at intervals of 65, 80, and 100 ms not significantly different from that for a single 3 ms pulse. Up to 50 ms, the difference in probability between single and paired stimuli was not significant, illustrating a dominant transient contribution. When the second transient was no longer effective, the sustained portion of the stimulus maintained a high response probability. However, the

Interpretation of these probability data may be influenced by a ceiling effect for single stimuli.


Laten"

Response latency was affected by increasing stimulus duration or interpulse interval, F- (6,90)=2.22, A<.05, and stimulus type had a significant effect, E (1,15)=10.40, pL<.01, but the stimulus type by time interaction was not significant (see Figure 3). For single stimuli, Increasing duration had an effect on latency, E (6,90)=4.43, .R<.001, but this was due to the increase from 3 to 20 ms only, since, if the 3 ms data were not included in the analysis, no duration effect was found (E (5,75)=0.64). For paired stimuli, increasing interpulse interval had no effect on response latency.



Discussion

The findings of Experiment 1 show that temporal summation of transient acoustic input dominates that of sustained input, In adults. Transients occurring within a window of 35-50 ms summate, and sustained


















I/
/
/


- PAIRED "SINGLE


I 2 5 20 355


50 65 80


DURATION/ INTERVAL (ms)

Figure 3. Adult response latency as a function of time for single and paired stimuli.


62 60


58 56



54 52


:3


I
100







24

summation adds little which cannot be explained by transient summation In this range. Beyond this window, transient summation does not occur, and the contribution of sustained summation becomes apparent. Also, these differential summation effects are reflected by startle response amplitude, probability, and latency in different ways.

For response amplitude the summation window for transients is 3550 ms, and the time between transients has a slight, but not significant, effect on the response. As interpulse Interval within the window increases, a slight increase In response amplitude may be explained by forward masking, which suggests that the first pulse decreases the effectiveness of the second. As the interpulse interval Increases the amount of masking decreases (Raab, 1961), so that responding at the 20 and 35 ms intervals may differ due to a reduction in responding at the 20 ms interval, but not at the 35 ms interval (or, at least, not as great a reduction). If forward masking is present, its Influence is not great, possibly due to the brief duration of the first pulse (Zwislocki, 1978). Also, this marginal masking effect was found for response amplitude, but not for response probability.

For response probability, the summation window extends to 50 ms, and the effect of the second transient Is not influenced by where in this window It occurs, above 20 ms. For response latency, the second transient had no apparent effect at any of the intervals used. It may be the case that the temporal summation window for transients Is narrower than 20 ms, so that the second pulse always occurred outside the window, and this prediction may be supported by future research. For sustained summation, 20 ms Is adequate, but the window may be less than this.









After the transient summation window has been exceeded, sustained summation continues to occur, either increasing responding or maintaining elevated responding after the effect of the second

transient has disappeared. This peaks at 50 ms for response amplitude, and at 20 ms for response latency. For response probability, the peak of the sustained summation function is difficult to specify due to the presence of a ceiling effect at high probabilities, but this peak could not exceed the peak response latency (approximately 65 ms). Alsoffl the extent to which the second pulse in a pair elicited a response when the first pulse failed to do so was minimal (less than 10 instances for all subjects combined). This means that the second pulse alone is not effective in eliciting responding.

The temporal summation of transient stimuli peaked at 35 ms for response amplitude and 50 ms for response probability. Within these ranges, the contribution of the sustained portion of a stimulus was no greater than that of the second brief transient. Beyond these ranges, the second stimulus in a pulse pair had little or no effect, but responding was maintained by the sustained portion of a single stimulus. A second experiment was conducted, to investigate these temporal summation properties in greater detail.















CHAPTER III
EXPERIMENT 2

Introduction

The data of Experiment 1 suggest that the temporal summation window for transient stimuli Iis 35-50 ms for response ampliItude and probability, and less than 20 ms for response latency. To more accurately assess respond ing in the critical range of 35-50 ins, a second experiment was conducted, in which stimulus duration and interpulse intervals in the 30-55 ms range were varied in 5 ms steps.

Method

Sub4lets

Subjects were 12 university students, 6 males and 6 females, with an average age of 22 years (range=18 to 29), selected in the same way as in Experiment 1.


Stimuli. Apparatus. Procedure, and Data Analysisj

These were identical to those used In Experiment 1, except that stimulus durations were 3, 30, 35, 40, 45, 50, and 55 ins, with analogous Tnterpulse Intervals for paired (3 ins) stimuli.


Results

Amplitu~de

Response amplitude was affected by increasing stimulus duration or interstimulus interval, E- (6,66)= 3.70, p- <.005, and stimulus type had a significant effect, E- (1,11)= 29.71, p- <.001, but the two variables









did not interact (see Figure 4). For single stimuli, a significant linear trend for duration was found, E (1,11)= 9.70, p- <.01. A single 30 ms stimulus was more effective than a 3 ms stimulus, � (11)=2.86, p. <.025, but further Increases in stimulus duration were not significantly more effective than a 30 ms stimulus. For paired stimuli, a significant quadratic time trend was found, E- C1,11)= 6.81, P. <.025. Pairs of pulses at intervals of 35, 40, and 45 ms were slightly, but not significantly, more effective than a single 3 ms stimulus ( I (11)= 2.09, p. <.06; 1 (11)= 1.98, p- <.08; 1 (11)= 2.02, p- <.07, for 35, 40, and 45 ins, respectively). At 50 and 55 ms intervals, pairs of pulses were clearly no more effective than single 3 ins stimuli. Single stimuli were significantly more effective than paired stimuli at 50 ins, � (11)=

3.44, p- <.01, but not below.


Probability

The effect of Increasing stimulus duration or intersiinulus interval differed for response probability, as illustrated by a significant stimulus type by time Interaction, E (6,66)= 2.27, p. <.05 (see Figure 5). For single stimuli, a linear trend for duration was found, E (1,11)= 7.53, p. <.025. A 30-ms duration stimulus was more effective than a 3 ms stimulus, � (11)= 2.40, p. <.05, but no significant difference was found between a single 3 ms pulse and pairs of pulses at all other intervals tested. For paired stimuli, a significant quadratic interval trend was found, E (1,11)= 5.64, p. <.05. A pair of pulses 30 ms apart were more effective than a single 3 ins stimulus, � (11)= 2.52, p. <.05, but further increases in interstiinulus Interval had no significant effect. A single stimulus was more







































3 30 35 40 45 50 55

DURATION/ INTERVAL (ms)

Figure 4. Adult response amplitude as a function of time for single and paired stimuli.


45 40 35


/SINGLE










PAIRED


30 25


20





















SINGLE




PAIRED










3 20 25 30 35 40 45


100 90 80


60 50


DURATION / INTERVAL (ms)


Figure 5. Adult response probability as a function of time for single and paired stimuli.









effective than an analogous pair of stimuli at 40, 50, and 55 ms

(11)= 2.32, 2.28, and 2.54, respectively, p- <.05). Laten"

A significant stimulus type by time interaction was found for response latency, F- (6,66)= 2.54, p- <.05 (see Figure 6). For single stimuli, the overall time effect was not significant, but latency was lower for 40 ms stimuli than for 3 ms stimuli, � (11)= 3.39, p- <.01. For paired stimuli, a linear effect of interstimulus Interval was found ' E (1,11)= 14.65, p- <.005. Increasing interval from 30 to 50 ms resulted In a significant increase in response latency, � (11)= 2.75, A <.025, but no other comparisons of paired stimuli were significant. Latencies for single stimuli were significantly lower than those for paired stimuli, at each interval tested.



Discus '

For response amplitude, increasing stimulus duration to 30-55 ms resulted in larger responses, as illustrated by a significant linear trend for duration. Increasing Interstimulus interval for paired stimuli resulted In larger responses up to 40 ms, then a reduction In response amplitude, as shown by the significant quadratic function for paired stimuli. Single stimuli were not significantly more effective than paired stimuli below 50 ms. In Experiment 2, the two functions diverged between 35 and 50 ms, just as they did in Experiment 1. The fact that the response amplitude functions are so similar in the two experiments suggests that these findings are reliable. Up to 35 ms, the transient aspects of the stimulus seem almost as effective as the



















\ PAIRED


DURATION /INTERVAL (ms)

Figure 6. Adult response latency as a function of time for single and paired stimuli.


68 66

64 62


6oL


58 56

54


SINGLE


I I 1 1
30 35 40 45


1 1
50 55









sustained aspects in determining response amplitude.

In Experiment 1, it was shown that temporal summation for response probability may peak below 20 ms, and this was supported In Experiment 2. Increasing stimulus duration from 3 to 30 ms resulted in Increased response probability, with further increases to 55 ms having no greater effect. A pair of stimuli 30 ms apart were more effective than a single 3 ms pulse but further Interval Increases were not effective. Paired pulses were as effective as single pulses up to 45 ms, above which the effect of the second pulse in the pair tended to decrease. This parallels the finding in Experiment 1 that transient summation is as effective as sustained summation up to 50 ms.

The response latency data of Experiment 2 are also similar to those of Experiment 1, but not as clearly as were the amplitude and probability data. Latency was lower for single than for paired stimuli, in all cases. However, where the latency effect for single stimuli peaked at 20 ms in Experiment 1, it peaked at 40 ms In Experiment 2. Also, where no effect of latency was found for paired stimuli In Experiment 1, longer latency was found at 50 than at 30 ms Intervals in Experiment 2. The reason for these latency differences In the two experiments is not clear, but these data suggest that latency may be a less reliable measure of startle activity than either response

amplitude or probability.

Experiments 1 and 2 show that the temporal summation of acoustic stimuli occurs in different ways for transient and sustained stimuli, and that this differential summation is reflected in different ways by startle amplitude, probability, and latency. For response amplitude, transient summation dominates for 40 ms, and sustained summation adds









little during this time. Beyond the transient summation window, sustained summation continues to Increase response amplitude for a brief time (<20 ms). For response probability, summation of transients peaks at less than 20 ms, and continues to 50 ms, with minimal contribution by sustained summation in this period. Above 50 ms, sustained summation maintains elevated probability while the effect of a second transient disappears. The probability and amplitude functions are similar in that the temporal summation window for transients is not more than 50 ms in both cases, but this summation seems to peak earlier for probability than for amplitude. Response latency appears to reflect summation properties similar to those shown by probability, but In a less reliable fashion.

Experiments I and 2 illustrate the fact that the startle

elicitation paradigm provides an effective tool for comparing transient and sustained system activity. Therefore, it would be of interest to apply this paradigm to infant subjects, since Graham, Anthony, & Zeigler (1983) hypothesize that the transient system is relatively immature when compared to the sustained system in infants. Experiment 3 measured startle amplitude, probability, and latency to stimuli analogous to those used in Experiments 1 and 2. If transient immaturity exists, these measures should show differences between Infants and adults.














CHAPTER IV
EXPERIMENT 3

Method



All subjects were apparently healthy newborns, with no significant problems or complications, and were the result of spontaneous vaginal delivery at Shands Hospital in Gainesville, Florida. Data from 4 of 24 subjects tested were not used, due to a failure to respond to at least two trials at each stimulus condition. For the 20 remaining infants, average age was 23 hours (range=7-36 h), average gestational age by maturity rating was 39.75 weeks (range=37-42 wk), average 5 minute

Apgar (1953) score was 8.9 (range=8-10), and average birthweight was 3350 gm (range=2640-4220 gm). Testing began an average of 1 hour 10 minutes after the last feeding (range= 30 min- 3 h 30 min). Local anaesthetic was administered to 11 mothers during delivery, with no anaesthetic in the remaining 9 cases.




Stimuli were produced by a Grason-Stadler 455C noise generator, gated through a Coulbourn electronic switch and a Dynaco Stereo 120 amplifier, and presented through a JBL Decade 26 loudspeaker, located 1 m in front of the subjects. Sound levels were checked before each session, using a General Radio 1551C sound level meter and a Simpson 260 voltmeter. The EMG signal was amplified and integrated In the same way as in Experiment 1, except for some variation In Bioamplifier gain









settings. This signal was sampled by an Apple 11 Plus computer (12 bit accuracy) every millisecond for 250 ms and corrected for gain variations. EKG was recorded on a Teac A-2300SX tape recorder, and on a Narco DMP-4B3 Physiograph. Respiration was recorded using a Parks Electronics strain gauge and a Parks Electronics Model 270 plethysmograph, and this signal was also displayed on the Narco Phys lograph.


Stimuli

Stimuli were broadband noise bursts (2OHz-20 KHz) at 100 dBCA) with fast (< 0.1 ins) rise and fall times. Pilot data showed that 95 dB(A) stimuli (the intensity used In Experiments 1 and 2) did not reliably elicit responding at 3 ms duration. Five of the stimuli varied in duration (3, 20, 35, 50, and 100 ins), and the other four were pairs of 3 ms pulses at onset-to-onset intervals of 20, 35, 50, and 100 ins. Interstimulus interval ranged from 20-60 ins.




Each subject was placed supine In a crib in the testing chamber, and the respiration strain gauge was attached laterally across the abdomen, just above the navel. EKG electrodes were attached in a modified Lead 11 configuration. The area just below the left eye was then cleaned with a cotton swab soaked in alcohol, and the EMO electrodes were attached, one just below the center of the eye and the other immediately temporal to the first, as close to the orbital ridge as possible. Stimulus presentation was control led by an experimenter, sitting beside the loudspeaker, who also recorded the state of the Infant on each trial, using the rating scale of Berg, Berg, and Graham









(1971). These behavioral ratings, along with respiration and EKG measures, were used to determine the state of the infant, and these three measures have been shown to be adequate and reliable indicators of sleep state in young infants (Anders, Emde, & Parmelee, 1971). Only trials during which the Infant was judged to be in quiet sleep were included In the analysis. Stimuli were not presented during movement of the head, eyes, or arms, or when the subject's eyes were open. A session was terminated when 9 potentially scramble responses at each stimulus condition were obtained. If the Infant changed state the session was halted, and the infant was allowed to return to quiet sleep. If the subject did not reach quiet sleep within an hour, the truncated session was terminated. Stimulus order was determined by a 9x9 Latin square, with the row of entry into the square being randomly determined. Response amplitude, probability, and latency were measured in the same way as in Experiment 1.

Results

Amplitude

A significant effect of stimulus type was found, E (1,19)=5.13, .R<.05, with single stimuli eliciting larger responses than paired stimuli (see Figure 7). A significant time effect was also found, E (4,76)=6.46, p-<.001, but stimulus type and time did not interact. For single stimuli, increasing duration from 3 to 20 ms yielded larger responses, � (19)= 2.09, p.<.05, and a further increase from 20 to 35 ms resulted in still larger responses, � (19)=4.07, p-<.001. Increasing stimulus duration above 35 ms had no further effect. For paired stimuli, increasing interpulse interval from 20 to 50 ms resulted In larger responses, � (19)=2.12, p-<.05. Responding to a pair of pulses at









100


90 so


--- SING LE






PAIRED


60 50


1
20


1 1
35 50


100


DURATION /INTERVAL (ms)


Figure 7. Infant response amplitude as a function of time for single and paired stimuli.









an interval of 100 ms was not significantly different from that to a single pulse at 3 ms duration, suggesting that the second pulse in the pair did not contribute to the response.

A single stimulus 20 ms long was not significantly more effective than a pair of pulses 20 ms apart. At 35 ms, single stimuli were more effective than pairs of pulses, � (19)=2.47, pL<.025, but this advantage was not found at 50 or 100 ms. These findings suggest that a pair of pulses can be nearly as effective as an analogous single stimulus, up to 50 ms. The transient contribution seemed to be dominant In determining responding, with the sustained portion of the stimulus adding little.


Probability

For response probability, an Interaction between stimulus type and time was found, E (4,76)=20.67, p-<.001 (see Figure 8). For single stimuli, a quadratic duration trend appeared, E (1,19)=24.28, PL<.001. Response probability increased as stimulus duration was Increased from 3 to 20 ms, I (19)=2.76, A<.025, and from 20 to 35 ms, � (19)=4.41, ,Q<.001, with duration increases above 35 ms having little effect. For paired stimuli, a quadratic interval effect appeared, E (1,19)= 19.79, A<.001. Probability was not affected by Increasing interpulse interval to 50 ms. At 100 ms interval, probability returned to the level found with a single 3 ms pulse, suggesting that the second pulse at 100 ms did not contribute to responding.

A pair of pulses at 20 ms Interval was more effective than a single 3 ms pulse, � (19)=2.82, p-<.01, and no less effective than a 20 ms long stimulus. Single stimuli were more effective than paired pulses





100


- SINGLE













PAIRED


I


90 80 70 60 50


I I I


100


3 20 35 50


DURATION/ INTERVAL (ms)

Figure 8. Infant response probability as a function of time for single and paired stimuli.









at 35, 50, and 100 ms, indicating a significant contribution of sustained stimulus attributes above 20 ms. Laten"

A significant effect of stimulus type on response latency was found., E (1,19)=4.53, p-<.05, with latency being slightly longer for paired than for single stimuli (except at 100 ms) (see Figure 9). No significant time effect or stimulus type by time interaction was found. Pairwise comparisons, looking at time comparisons for each stimulus type, and stimulus type comparisons for each time, showed that only the 50 ms single and paired stimuli resulted in different response latency, � (19)=2.42, R<.05. These data are difficult to interpret, due to the high latency variability, an effect also reported by Zeigler (1978) in 6-to-9-week-old Infants.


Discussion

The present study suggests that the temporal summation of auditory stimuli in human newborns follows a different time course for transient and sustained stimuli. Also, startle amplitude, probability, and latency do not reflect this differential summation in the same way.

Response amplitude increases if a second transient is added up to 50 ms after the first. In general, the sustained portion of the stimulus does not seem to contribute much that cannot be accounted for by the transient components of the single stimuli. This suggests that the window for temporal summation of transients is 50 ms for response amplitude, and that transient summation dominates sustained summation in this range.


















60



58



~56

z
54\ PAIRED



z
0
0
U)
E 50 SNL






3 20 35 50 100
DURATION /INTERVAL (inS)


Figure 9. Infant response latency as a function of time for single and paired stimuli.







42

The fact that response amplitude increases as interpulse interval increases to 50 ms suggests that forward masking may be present, with the first pulse in the pair decreasing the response to the second pulse. As interpulse interval increases, masking decreases (Raab, 1961), allowing the second pulse to contribute more to the response.

The findings for response probability are quite different. Adding a second transient 20-50 ms after the first results in Increased likilhood of responding, and the effect of the second transient is not influenced by where it occurs within this window. Sustained summation becomes Important above 20 ms, resulting in an increase in probability which asymptotes at 55 ms. This suggests that the window for temporal summation of transients Is 20-50 ms, and that sustained summation above 20 ms can add to this, with this sustained contribution reaching its maximum within 35-50 ms of stimulus onset. The effect of the second transient is less pronounced for probability than for response

amplitude, so transient and sustained system activity coexist, and both contribute significantly to responding In infants.
















CHAPTER V
GENERAL DISCUSSION


The startle response Is influenced by activity in both the transient and sustained systems, a finding contrary to the hypothesis of Dykman and Ison (1979). These two systems are at least partially independent, with the startle response being the net result of activity in both systems. These experiments show that manipulation of the startle-eliciting stimuli can demonstrate Independent Influences of transient and sustained system activity. However, the relative contribution to the startle response of these two systems may not be equal, since these results suggest that the transient system Is more efficient than the sustained system. As one system becomes dominant over the other, their relative contributions to the response will change.

To find the relative contribution of transient and sustained activity, it is first necessary to calculate the effect of a second 3 ms pulse in a pair if no transient activity were to occur for that pulse. This is accomplished by calculating the slope of the single stimulus function (in units per ms), and adding 3 times this value to the response to a single 3 ms stimulus. If responding to two pulses is simply additive, then 6 ms of stimulation should be equivalent to two 3 ms stimuli. The extent to which the second pulse in a pair affects responding above this predicted level indicates the influence of transient system activity.







44

The ratio of actual responding to predicted responding (based on linear energy integration) to the second pulse Illustrates the relative dominance of transient or sustained system activity. If this ratio is greater than 1.0, transient activity is dominant, since the actual data exceed the predicted values. Ratios less than 1.0 illustrate sustained system dominance.

For response amplitude, the ratios of actual to predicted (linear) response changes due to the presence of a second pulse in a pair are 7.28, 6.13, and 8.72, for Experiments 1, 2, and 3, respectively. For response probability, these ratios are 5.33, 5.87, and 5.23, for Experiments 1, 2, and 3, respectively. Clearly, the effect of the second pulse In a pair Is much greater than would be expected if only linear energy integration were occurring. The transient properties of the second pulse enhance responding, and transient system activity dominates sustained system activity when the two occur together.

The response amplitude data suggest that the result of temporal summation of transients is more efficient than that of summation of sustained stimulus attributes, for both neonates and adults. However, the window of temporal summation is narrower in adults (40 ms) than in infants (50 ms). Beyond this window, sustained summation becomes evident. Research using longer durations and intervals Is necessary to more accurately specify the transient summation window for neonates.

The infant and adult response probability data suggest that transient summation is maximal within 20 ms, and extends to at least 50 ms, with infant probability being about 20 percent lower than adult probability (see Figure 10), despite the fact that the stimulus was more Intense for infants than for adults. The temporal summation of

























ADULT I N FANT


9080 70 60 50

'1',


3 20 35 50 65 80


100


INTERVAL (ms)


Figure 10. Adult and infant response probability as a function of time for paired stimuli.


100


I I I


I 8i I Ii









transients follows the same time course in adults and neonates, for response probability. For adults, sustained summation contributed little more than could be accounted for by transient summation, and this may have been due to a ceiling effect. In infants, where maximal temporal summation of transients still yielded probabilities below 70 percent, a clear additive effect of sustained summation was found above 20 ins. It may be the case that sustained summation would be more apparent in adults if stimuli which avoid the response probability ceiling were used. When comparing infant and adult probability for single stimuli, the most striking finding is the much lower response probability for very brief stimuli in infants (see Figure 11). Adults require much less stimulation than do infants to achieve equivalent response probabilities. However, at longer durations single stimuli are quite effective in infants, supporting the belief that the sustained system is relatively mature In infants (Graham, Strock, & Zeigler, 1981).

Graham, Strock, and Zeigler (1981) suggest that transient immaturity may be due to both motor system and sensory system immaturity. The response of the orbicularis oculi, the muscle responsible for the eyeblink, to a startle stimulus consists of two components, R1 and R2, which are neurologically dissimilar (Sanes, Foss, & Ison, 1982; Shahanl & Young, 1973). The R1 response is a short latency (5-15 ins) response of motoneurons found in the palpebral segment of orbicularis oculi. This R1 component occurs Ipsi laterally to electrical and tactile, but not auditory, stimuli. The R2 component has a longer latency (20-100 ins), and occurs in the orbital segment of orbicularis oculi bilateral ly. The neural circuit for R2 is more















100


--- ADULTS


-INFANTS


















I
100


90 80 706050




3


I I I I I 20 35 50 65 80


(M S)


DURATION


Figure 11. Adult and infant response probability as a function of time for single stimuli.









complicated than that for R1, involving brain stem nuclei at the level of the Inferior colliculus. This R2 component is what is measured as the startle eyeblink by a periorbital electrode placement. For this reason, the R1 component may be Irrelevant to the present study. This means that the distinction between transient and sustained motoneurons may not be a relevant factor In defining the immaturity of the transient system, contrary to the suggestion of Graham, Strock, & Zeigler (1981). On the other hand, the suggestion by Graham, Strock, & Zeigler (1981) of a neurological dichotomy of transient and sustained sensory neurons is supported by the present research. The fact that pairs of stimuli can be as effective as analogous single stimuli, which contain much more energy, suggests that the input to the startle center may be equivalent for the two types of stimuli. This means that temporal summation occurs either before or at the startle center, due to differential activation of transient and sustained sensory systems.

In a previous report (Blumenthal & Berg, in press), it was suggested that startle amplitude and probability may be determined by partially independent underlying mechanisms, a startle trigger and a startle amplifier. This Is analogous to the distinction made in cognitive perceptual research between stimulus detection and stimulus identification (Posner, 1978).

The extent to which the two stimuli in a pair of transients have common neural elements may have an effect on their summation. If the units have not fully recovered from the effects of the first pulse, then the contribution of the second pulse may be decreased. This is a type of forward masking, and the recovery of these units may be immature in neonates. This would explain the finding that response









amplitude increases as interstimulus interval increases above 20 ms for infants, but not for adults. This neural recovery may be more extended for the startle amplifier than for the trigger, since no evidence of increase In forward masking was found over the 20-50 ms range of Intervals tested, for either Infant or adult response probability.

The exact location and nature of the trigger and amplifier mechanisms is unknown, but they are most likely to be found either in the auditory pathway below the inferior colliculus, or In the startle center itself. Also, the two mechanisms may have different locations, since they appear to be partially independent in their responding. If either the trigger or the amplifier is located in the auditory pathway, similar mechanisms may be found in other sensory pathways, and further studies investigating startle trigger and amplifier mechanisms in visual and tactile modalities are suggested. The fact that analogous temporal summation abilities are found across modalities suggests that this summation may be a very useful means of investigating startle trigger and amplifier mechanisms cross-modally. It would also be useful to conduct comparative studies using animals, to more accurately locate the mechanisms underlying these differential findings. Such studies might show that the trigger and amplifier mechanisms are due to specific startle center involvement or to basic organizing principles found In all sensory systems.

In adults, sensitivity to transients may be greater for the trigger than for the amplifier, since response probability peaks at 20 ms for both single and paired stimuli, while the amplifier continues to summate until 40 ms. This implies that the amount of energy required to reach maximal activation is lower for the trigger than for the









amplifier. In fact, this critical level may be below 20 ms, and further research at durations and intervals of less than 20 ms is needed to measure the transient sensitivity of the startle trigger.

In infants, the immature transient sensitivity of the startle trigger is illustrated by the low response probabilities for brief duration stimuli (see Figure 11). The immaturity of the startle trigger is further illustrated by the fact that pairs of pulses are much less effective in infants than In adults (see Figure 10). However, the infant and adult curves are parallel, so it may be the case that interstimulus interval is not the crucial parameter. The probability data suggest that both the single and paired stimulus functions show age differences which are due to immature responding to very brief stimuli, regardless of the interstimulus interval. Making individual pulses in each pair longer, so that the critical energy could be reached at the shortest interval, Is necessary to separate the effects of increasing stimulus duration and Interstimulus interval. Future research using pairs of longer stimuli is suggested, to separate these duration and interstimulus Interval effects more fully.

The present data support Graham and coworkers (1981) in their belief that the transient system is immature in young infants. The infant response is immature In three ways. First, the window of temporal summation of transients, as illustrated by response amplitude, Is wider in infants than in adults. Second, the degree to which brief stimuli elicit the startle response, as illustrated by response probability, is more pronounced in adults than In infants. Third, pairs of brief stimuli are less effective in eliciting transient system activation at all intervals In infants than in adults, allowing






51

transient and sustained summation to occur together in Infants, for response probability.
















REFERENE


Anders, T., Emde, R., & Parmelee, A. A. (1971) A Manual of Standardized
Terminology. Techniques, and Criteria for Scoring of States of Slee 'p
and Wakefulness in Newborn Infants. UCLA Brain Information Service, Los
Angeles, CA.

Apgar, V. (1953) Proposal for a new method of evaluation of the newborn
Infant. Anesthesia and Analgesia, 3Z, 260-267.

Bartley, S. H. (1969) Principles of Perception, 2nd EditioLn. New York:
Harper & Row.

Berg, K. M. (1973) Elicitation of acoustic startle in the human. (Doctoral
dissertation, University of Wisconsin, 1973) Dissertation Abstracts
International, 3A, 5217B3-5218B.

Berg, K. M., Berg, W. K., & Graham, F. K. (1971) Infant heart rate
response as a function of stimulus and state. Psychophysiology, B, 3044.

Blumenthal, T. D., & Berg, W. K. (1982) Developmental differences in
startle response to stimulus onset transition. Psychophysiology, 12,
551-552 (abstract).

Blumenthal, T. D., & Berg, W. K. (1984) Differential effects of transient
and sustained components of acoustic stimuli on startle blink.
Psychophysiology, 21, 569 (abstract).

Blumenthal, T. D., & Berg, W. K. (in press) Stimulus rise time, intensity,
and bandwidth effects on acoustic startle amplitude and probability.
Psychophysiology.

Bohlin, G., Graham, F. K., Silverstein, L. D., & Hackley, S. A. (1981)
Cardiac orienting and startle blink modification in novel and signal
situations. Psychophysiology, J-6, 603-611.

Brown, J. H. (1966) Magnitude estimation of angular velocity during
passive rotation. Journal of Experimental Psychology, 22, 169-172.

Brown, J. H. (1968) Cross-modal estimation of angular velocity. erception
and Psychophysics, 3, 115-117.

Davis, M., & Gendelman, P. M. (1977) Plasticity of the acoustic startle
response in the acutely decerebrate rat. Journal of Comparative and
Physiological Psychology, 91, 549-563.









Davis, M., Gendelman, D. S., Tischler, M. D., & Gendelman, P. M. (1982) A
primary acoustic startle circuit: Lesion and stimulation studies.
Journal of Neuroscience, 2, 791-805.

Davis, M., & Heninger, G. R. (1972) Comparison of response plasticity
between the eyeblink and vertex potential in humans.
Electroencephalography and Clinical Neurophysiology, 33, 283-293.

Dykman, B. M., & Ison, R. (1979) Temporal integration of acoustic
stimulation obtained in reflex inhibition in rats and humans.
Journal of Comparative and Physiological Psychology, 93, 939-945.

Fleshler, M. (1965) Adequate acoustic stimulus for startle reaction In the
rat. Journal of Comparative and Physiological Psychology, 60, 200-207.

Fox, J. E. (1979) Habituation and pre-stimulus inhibition of the auditory
startle reflex in decerebrate rats. Physiology and Behavior, 23, 291297.

Gallager, D. W., Kehne, J. H., Wakeman, E. A., & Davis, M. (1983)
Developmental changes in pharmacological responsivity of the acoustic startle reflex: Effects of picrotoxin. Psychopharmacology, 19, 87-93.

Gersuni, G. V. (1971) Temporal organization of the auditory function. In
G. V. Gersuni (Ed.), Sensory Processes at the Neuronal and
Behavioral Levels. New York: Academic Press.

Geschelder, G. A. (1976) Evidence in support of the duplex theory of
mechanoreception. Sensory Processes, 1, 68-76.

Graham, F. K. (1975) The more or less startling effects of weak
prestimulation. Psychophysiology, 12, 238-248.

Graham, F. K. (1979) Distinguishing among orienting, defense, and startle
reflexes. In H. D. Kimmel, E. H. van Olst, and J. F. Orlebeke (Eds.),
The Orienting Reflex in Humans. Hillsdale, NJ: Lawrence Erlbaum
Associates.

Graham, F. K., Anthony, B. J., & Zeigler, B. L. (1983) The orienting
response and developmental processes. In D. Siddle (Ed.), Ihe
Orienting Response. New York: J. Wiley and Sons.

Graham, F. K., & Murray, G. M. (1977) Discordant effects of weak
prestimulation on magnitude and latency of the reflex blink.
Physiological Psychology, 5-, 108-114.

Graham, F. K., Strock, B. D., & Zeigler, B. L. (1981) Excitatory and
inhibitory influences on reflex responsiveness. In W. A. Collins (Ed.),
Aspects of the Development of Competence. Hillsdale, NJ: Lawrence
Erlbaum Associates.

Green, D. M. (1973) Minimum integration time. In A. R. Moller (Ed.), Basic
Mechanisms In Hearino. New York: Academic Press.









Harbin, T. J., & Berg, W. K. (1983) The effects of age and prestimulus
duration upon reflex inhibition. Psychophysiology, 2, 603-610.

Hickey, T. L. (1977) Postnatal development of the human lateral geniculate
nucleus: Relationship to a critical period for the visual system.
Science, 198, 836-838.

Higashiyama, A., & Tashiro, T. (1983) Temporal and spatial integration for
electrocutaneous stimulation. Perception and Psychophysics, 33, 437442.

Hirsch, H. V. B., & Leventhal, A. G. (1978) Functional modification of the
developing visual system. In M. Jacobson (Ed.), Handbook of Sensory
Physiology, V.9: Development of Sensory Systems. New York: SpringerVerlag.

Hoffman, H. S., & Ison, J. R. (1980) Reflex modification in the domain of
startle: 1. Some empirical findings and their implications for how the nervous stystem processes sensory input. Psychological Review, d7, 175189.

Kletsky, E. J., & Stengrevics, J. M. (1972) Estimation of the integration
time-constant In auditory receptor units. Journal of the Acoustical
Society of America, 52, 141 (abstract).

Lacey, B. C., & Lacey, J. I. (1974) Studies of heart rate and other bodily
processes in sensorimotor behavior. In P. A. Obrist, A. H. Black, J.
Brener, and L. V. DiCara (Eds.), Cardiovascular Psychophysiology:
Current Issues in Response Mechanisms.Biofeedback and Methodology.
Chicago, IL: Aldine.
Landis, C., & Hunt, W. A. (1939) The Startle Pattern. New York: Farrar and
Rinehart.

Marks, L. E. (1974) Sensory Processes. New York: Academic Press.

Marsh, R., Hoffman, H. S., & Stitt, C. L. (1973) Temporal integration in
the acoustic startle reflex of the rat. Journal of Comparative and
Physiological Psychology, B2, 507-511.

Norman, J. L., Pettigrew, J. D., & Daniels, J. D. (1977) Early development
of X-cells in kitten lateral geniculate nucleus. Science, 198, 202-204.

Posner, M. I. (1978) Chronometric Explanations of Mind. Hillsdale, NJ:
Lawrence Erlbaum Associates.

Prokasy, W. F., & Ebel, H. C. (1967) Three components of the classically
conditioned GSR in human subjects. Journal of Experimental Psychology,
73, 247-256.

Raab, D. H. (1961) Forward and backward masking between acoustic clicks.
Journal of the Acoustical Society of America, 31, 137-139.









Radionova, E. A. (1971) Two types of neurons in the cat cochlear nuclei
and their role in audition. In G. V. Gersuni (Ed.), Snsry
Processes at the Neuronal and Behavioral Levels. New York: Academic
Press.

Richards, A. M. (1976) Basic Experimentation in Psychoacoustics.
Baltimore, MD: University Park Press.

Sanes, J. N., Foss, J. A., & Ison, J. R. (1982) Conditions that affect the
thresholds of the components of the eyeblink reflex in humans. Journal
of Neurology. Neurosurgery. and Psychiatry, 45, 543-549.

Scharf, B. (1978) Loudness. In E. C. Carterette and M. P. Friedman (Eds.),
Handbook of Perception. V.4: Hearing. New York: Academic Press.

Schwartz, S. H., & Loop, M. S. (1984) Effect of duration on detection by
the chromatic and achromatic systems. Perception and Psychophysics, 3,
65-67.

Shahani, B. T., & Young, R. R. (1973) Blink reflexes In orbicularis oculi.
In J. E. Desmedt (Ed.), New Developments In Electromyography and Clinical Neurophysiology, Vol. 3. Basel, Switzerland: S. Karger.

Small, A. M., Jr., Brandt, J. F., & Cox, P. G. (1962) Loudness as a
function of signal duration. Journal of the Acoustical Society of
America, 3A, 513-514.

Stevens, J. C., & Hall, J. W. (1966) Brightness and loudness as functions
of stimulus duration. Perception and Psychophysics, _, 319-327.

Stevens, S. S. (1966) Duration, luminance, and the brightness exponent.
Perception and Psychophysics, 1, 96-100.

Stitt, C. L., Hoffman, H. S., & Marsh, R. R. (1976) Interaction versus
independence of startle-modification processes in the rat. Jurnalf.
Experimental Psychology: Animal Behavior Processes, 2, 260-265.

Verrillo, R. T. (1968) A duplex mechanism of mechanoreception. In D. R.
Kenshalo (Ed.), The Skin Senses. Springfield, IL: Charles C. Thomas
Pub.

Yamada, F. (1983) Effects of intensity, duration, and rise and fall times
of the reflex evoking sound upon auditory evoked EMG of orbicularis
oculi and subjective estimation of startle. Japanese Journal of
Physiological Psychology and Psychophysiology, 1, 11-18.

Zeigler, B. L. (1978) Acoustic-startle modification by sustained and
transient components of long lead stimuli in human infants and adults.
Unpublished Master's thesis, University of Wisconsin - Madison.










Zeigler, B. L., Strock, B. D., and Graham, F. K. (1979) Facilitatory and
inhibitory reflex preconditioning in early infancy. Paper presented at the Society for Research in Child Development Meetings, San Francisco,
March, 1979.

Zwislocki, J. J. (1960) Theory of temporal auditory summation. Journal
of the Acoustical Society of America, 32, 1046-1060.

Zwlslocki, J. J. (1969) Temporal summation of loudness: An analysis.
Jour Lalof the Acoustical Society of America, 46., 431-441.

Zwislocki, J. J. (1978) Masking: Experimental and theoretical aspects of
simultaneous, forward, backward, and central masking. In E. C.
Carterette and M. P. Friedman (Eds.), Handbook of Perception,
.4Haing. New York: Academic Press.

Zwlslocki, J. J., & Sokolich, W. G. (1974) On loudness enhancement of a
tone burst by a preceding tone burst. Perception and Psychophysics, 16.,
87-90.














BIOGRAPHICAL SKETCH

Terry D. Blumenthal was born in Edmonton, Alberta, Canada, on July 17, 1954. He grew up In and around Roily View, Alberta, and eventually obtained a Bachelor of Science degree, with a specialization In psychology, at the University of Alberta in Edmonton. With this feat behind him, he decided to move from the top of a large pile to the bottom of a smaller pile, and he applied for graduate school. The University of Florida retained his services for five years, during which time he enjoyed life to the fullest, with considerable help from his friends and fellow graduate students in a variety of departments. While at the University of Florida, he worked under the supervision of Dr. W. Keith Berg, who had a profound influence on how Mr. Blumenthal viewed the world. Five years after his arrival in Gainesville, Mr. Blumenthal moved on to a teaching position at Hamilton College in upstate New York, where he taught courses, wrote his doctoral dissertation, and had a very productive year on all fronts. At this writing, matrimony looms on the horizon. Mr. Blumenthal has been consistently reminded of the considerable rewards which follow hard work, personal application, and a well-told joke.









I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




W. Keith Berg, Chair
Associate Professor of Psychology



I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




Kenneth J Gerhardt
Associate rofessor of Speech



I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.





Gra ,ate Reslearc ro essor of linical Psy ology



I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




Patricia H. Mil Ier
Associate Professor of Psychology








I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




oFnald .Ta
Professor of Psychology



This dissertati -on was submitted to the Graduate Faculty of the Department
of Psychology in the College of Liberal Arts and Sciences and to the Graduate School, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy.

August 1985


Dean, Graduate School





















































UNIVERSITY OF FLORIDA

3 1262 08285 260 8




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EK2T4NI6C_9AYZPK INGEST_TIME 2017-07-20T21:19:52Z PACKAGE UF00102784_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

DEVELOPMENTAL DIFFERENCES IN THE TEMPORAL SUMMATION OF TRANSIENT AND SUSTAINED AUDITORY STIMULI By TERRY D. BLUMENTHAL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1985

PAGE 2

Dedicated to my parents, Et ta Dora Blumenthal and Otto John Blumenthal, tor giving me the tools I need to get the job done.

PAGE 3

ACKNOWLEDGEMENTS I would I Ike to thank the members of my doctoral committee, Ors. W. Keith Berg (Chair), Kenneth J. Gerhardt, Peter J. Lang, Patricia H. Miller, and Donald C. Teas, for their help In bringing this work into being and for their generous donations of time and ideas. It is always helpful to have interested and learned experts to provide comments and critique. would further I Ike to thank Dr. Berg for his unfall Ing friendship and guidance throughout my graduate career. His constant and energetic drive to achieve qual lty and precision In research and teaching have been the ideals I compare my own endeavors to, In order to improve my ski I Is. l would also I ike to thank Kathy Berg for her helpful advice, and for writing such an Interesting dissertation. would I ike to thank Margarete Davies, Barry Hurwitz, Kristen Bomas, and Alonso Avendano for al I their help in the gathering and analysis of the data. I would further l Ike to thank Alonso Avendano for his patience and hard work beyond the cal I of duty. If not for his help, I would probably still be trying to collect these data. would also I ike to thank al I of the undergraduates and other adult subjects, the staff of the Newborn Nursery at Shands Hospital In Galnesvil le, Florida, and the parents and their new bundles of joy, for their patience and cooperation. am grateful to the students, faculty, and staff of the Department of Psychology at the University of Florida, for making iii

PAGE 4

graduate school such an Intensely enjoyable experience. am also grateful to the faculty, staff, and students of Hamilton College for showing me that there Is I ife after graduate school. Finally, would I ike to thank Dr. Berg again, for setting such a good example. iv

PAGE 5

TABLE OF CONTENTS ACKNOWLEDGEMENTS.......................................... ii I ABSTRACT. . . . v 1 CHAPTER I I NTRODUCT I ON Transient and Sustained Systems................... 1 The Startle Response.............................. 6 I I I I I . IV. v. EXPERIMENT I Method Res u I ts Dlscussion EXPERIMENT I I Introduction ...•...•.•....•...•••...•...........•• Method Results Discussion EXPERIMENT J 11 Met hod Results D i sc us s ion GENERAL DI SCUSS I ON REFERENCES .............................. BIOGRAPHICAL SKETCH 15 15 18 22 26 26 26 26 30 34 34 36 40 43 52 57

PAGE 6

Abstract of Dissertation Presented to the Graduate School of the University of Florida In Partial Fulf ii lment of the Requirements for the degree of Doctor of Philosophy DEVELOPMENTAL DIFFERENCES IN THE TEMPORAL SUMMATION OF TRANSIENT AND SUSTAINED AUDITORY STIMULI By Terry D. Blumenthal August 1985 Chairman: W. Keith Berg Major Department: Psychology The present study measured the startle eyebl ink electromyographic (EMG) response to transient and sustained auditory stimul I in human adults and newborns, to assess immaturity of temporal summation In newborns. Stlmul I were broadband noise bursts (20Hz-20KHz) with fast (<0.1 ms) rise times. In Experiments 1 and 2 adults received 95 dB(A) stimul I, in 13 stimulus conditions. One stimulus was 3 ms long, six varied In duration from 20 to 100 ms In Experiment 1 and from 30 to 55 ms in Experiment 2, and six consisted of pairs of 3 ms stimuli at onset-toonset intervals analogous to the durations used. In Experiment 3, newborn Infants (7-36 hours old) received 100 dB(A) stimul I analogous to those of Experiments 1 and 2. Startle amp I itude Increased for single and paired stlmul I up to 40 ms in adults and 50 ms In infants, above which single stimul I maintained this response amp I ltude while the second pulse in a pair was not effective. Within these summation windows, single stlmul i were not significantly more vi

PAGE 7

effective than analogous paired stlmul i. Startle probabll ity Increased for single stimuli up to 20 ms In adults and 35 ms in infants, above which no significant age difference was found. Single and paired stimuli were equally effective up to 45 ms for adults and 20 ms for Infants. For paired stimuli, intervals of 20-45 ms were equally effective in adults, and Intervals of 20-50 ms were equally effective in infants, with the second pulse at longer Intervals having no effect, but probability for Infants was lower than that for adults at al I intervals tested. Response latency decreased as stimulus duration increased to 20 ms, for adults but not for Infants, and no significant effect of interpulse Interval on latency was found at either age. The window for transient summation Is narrower In adults (40 ms) than In infants (50 ms), and brief stlmul i are more effective in adults than in infants. These data suggest that the transient system Is immature In Infants and that these developmental differences are II lustrated In different ways by startle amp I itude, probabll ity, and latency. vii

PAGE 8

CHAPTER I INTRODUCTION Transient and Sustained Systems The processing of sensory input may occur in different ways for transient and sustained stlmul I. Transient stlmul I activate onsetsensitive response systems (detection), while sustained stlmu 11 activate response systems which are sensitive to more prolonged stimulus attributes (identification). A stimulus with both transient and sustained properties may activate both types of response system. This transient-sustained dichotomy has been found for auditory (Graham, 1979), tactile (Verrl I lo, 1968), and visual (Schwartz & Loop, 1984) Input, and underlying neurological mechanisms have been proposed in each modal lty (Gersuni, 1971; Gescheider, 1976; Hickey, 1977). Also, the mechanisms underlying this dichotomy are believed to develop at different rates, at least in the auditory (Graham, Anthony, & Zeigler, 1983) and visual systems (Hickey, 1977; Norman, Pettigrew, & Daniels, 1977). One way in which transient and sustained systems differ Is In their temporal summation abil !ties, with rapid summation occurring In transient systems, and longer-lasting summation occurring In sustained systems. In the auditory system, temporal summation has been Investigated using behavioral seal Ing (Zwlslockl, 1969), single unit responding (Gersuni, 1971), and bralnstem reflex el icitatlon (Marsh, Hoffman, & Stitt, 1973). The present study used the acoustic startle 1

PAGE 9

2 reflex to assess developmental differences in temporal summation In human transient and sustained systems. Temporal summation involves the integration of sensory input over time. This is ii lustrated by the fact that the effectiveness of a stimulus can be increased by increasing its duration, to a point. The duration above which no further summation occurs is cal led the critical duration, and is the point at which the summation function (perception plotted as a function of stimulus duration) reaches asymptote. Also, a pair of stimuli can be more effective than a single stimulus, and this increases as interstimulus Interval decreases. The interval below which further reductions of interstlmulus interval have no greater effect can be considered a critical interval, analogous to the critical duration. Both the critical duration and the critical interval are the points at which their respective functions of sensation plotted as a function of time reach asymptote. Temporal summation has been found in several sensory modal itles, using a variety of experimental techniques. Psychophysical Investigations of visual perception show that brightness Increases as stimulus duration increases to 500 ms, but not beyond (Stevens, 1966). As stimulus intensity increases, this critical duration decreases, but the slopes of the summation functions are parallel across a range of intensities. This suggests that summation occurs unti I some specific amount of Input has occurred, and this amount is reached more rapidly by stimuli of higher Intensity. Temporal summation may asymptote due to sufficient "area under the curve", as determined by the stimulus envelope. Schwartz and Loop (1984) have shown that visual perception may involve two partially independent systems, a chromatic and an

PAGE 10

3 achromatic system. The chromatic system is sensitive to stimulus duration, showing summation functions with critical durations of 500 ms. The achromatic system is bel leved to be sensitive to transient stimulation only, showing no effect of duration In the range tested (43-1000 ms). These systems may be differentiated at the neuronal level, in that retinal ganglion eel ls with transient discharge patterns (Y-cel Is) are more common In the periphery, where achromatic stimulation is more pronounced (Hirsch & Levanthal, 1978). Gangl Ion eel Is with sustained discharge patterns (X-cel Is) are more common In and around the fovea, where chromatic activation is maximal. This indicates that the differential temporal summation of chromatic (sustained) and achromatic (transient) systems may be due to underlying neurological distinctions. For intermittent visual stimuli, brightness decreases as interstimulus Interval increases beyond the critical fl lcker fusion threshold (Bartley, 1969). Below this threshold, temporal summation occurs to prevent the brightness from decreasing when the stimulus Is not on continuously. In the visual system, the critical Interval is II lustrated by the critical flicker fusion threshold, Temporal summation has also been shown for both taste perception (Bujas & Ostojcic, 1939, as cited in Marks, 1974), and proprioception (Brown, 1966, 1968). Brown rotated subjects, varying the duration and acceleration of the rotation. The perceived magnitude of the rotation increased as the duration of the rotation Increased, at a range of accelerations. As acceleration was increased, the effect of duration on perceived magnitude of rotation reached asymptote earlier, similar to

PAGE 11

4 the finding on the visual system that critical duration decreases as stimulus intensity increases. For tactile stlmul i, two patterns of temporal summation exist. If Paclnlan corpuscles are duration of 500-1000 ms present, summation (Gescheider, 1976; occurs with a critical Verrillo, 1968). This critical duration decreases as stimulus Intensity Increases, just as in the visual and proprioceptive systems. lf Pacinian corpuscles are not present, summation stll I occurs, but with a critical duration of less than ms (Higashlyama & Tashiro,1983). This is analogous to the rapidly summating achromatic (transient) and slowly summating chromatic (sustained) systems found in vision. Both visual and tactile systems are distinguished by functional differences at the periphery which result in qualitatively different stimulus processing. In the auditory system, psychophysical methods have been used extensively to II lustrate temporal summation. Loudness matching has been used to show that increasing stimulus duration results In decreased threshold, up to a critical duration of 200 ms (Scharf, 1978). Critical duration decreases as stimulus Intensity increases (Richards, 1976), but the slope of the summation function Is the same across a wide range of Intensities (Smal I, Brandt, & Cox, 1962; Stevens & Hal I, 1966), as Is the case for tactile, proprioceptive, and visual stlmul i. For pairs of brief pulses, threshold increases as lnterpulse interval increases, up to 200 ms (Scharf, 1978; Zwislocki, 1960). Using psychophysical methods, the critical duration and critical interval for auditory temporal summation appear to be the same. While the use of loudness matching in assessing temporal summation has provided much information, this technique is not without problems.

PAGE 12

5 Differences in subject criteria used In the matching tasks can result in variable findings. For example, when subjects are told to match on the basis of the loudness at the onset of a stimulus as opposed to the total loudness of the stimulus, differences in stimulus duration effects are found (Zwlslocki & Sokol ich, 1974). These problems could be avoided by using techniques which are less susceptible to cognitive Influences, responses. such as psychophysiological or neurophyslological The investigation of auditory temporal summation at the neurophysiologlcal level has led to the Identification of two types of neurons in the auditory pathway which differ In their characteristic discharge patterns to stlmul I (Gersuni, 1971; Radlonova, 1971). Short time constant (STC) units ii lustrate a transient discharge at stimulus onset (and at offset, In some cases), while long time constant (LTC) units discharge in a sustained fashion, throughout the duration of the stimulus. These units are also sensitive to different stimulus parameters, with LTC units being sensitive to stimulus duration, but not to rise time, wh I le the opposite is true of STC units. The two types of units show different patterns of temporal summation, with STC units having very brief critical duratlons(a few mil I !seconds), and LTC units having much longer critical durations. These units are found at a I I I eve Is of the auditory pathway, and may be organ I zed into functional systems resulting In differential processing of transient and sustained stimulus attributes. In this respect, the LTC and STC systems are analogous to Paclnlan and non-Paclnian systems in tactile processing, and to chromatic (X-cel I) and achromatic (Y-cel I) systems in visual processing. Al I three modal !ties ii lustrate distinctive

PAGE 13

6 transient and sustained systems which possess different temporal summation properties. Evidence for auditory temporal summation has also been found at the level of the receptor (hair eel I) in monkeys (Kletsky & Stengrevlcs, 1972). Period histograms of receptor activity, which are correlated with positive excursions of the stimulus waveform, decay more slowly than does the stimulus, with a lag of .1 to .2 ms. This imp I ies that some integration may be taking place at the hair eel I, resulting in a prolongation of the half-wave rectified receptor potential. Temporal summation appears to be a fundamental property of several sensory systems. This summation may vary with the stimulus envelope, unti I some maximal amount of summation has occurred, since critical duration decreases as stimulus intensity increases, in several sensory modalities. Also, in the tactile, visual, and auditory systems, differential summation of transient and sustained stimul I can be found, suggesting the possibility of analogous underlying mechanisms in the three systems. Investigations of such mechanisms would be aided by a response which is sensitive to smal I stimulus changes, and which can be found across a range of ages and species. The startle reflex is such a response. The Startle Response The use of the startle response as a means to assess temporal summation can be recommended for several reasons. This response is very reliable, shows considerable plasticity and sensitivity to stimulus conditions, and is mediated by a relatively simple neural circuit.

PAGE 14

7 Also, the startle response is time-locked to stimulus onset, and can be measured nonlnvasively. A wealth of developmental and comparative data exists, ii lustratlng the general lzabi I lty of the startle response across studies, subjects, and species. Landis and Hunt (1939) found that the eyebl Ink is the first component of the startle response to appear, and that this reflex eyebl ink is very rel table, resisting habituation. The reflex eyebl ink Is also a sensitive measure, In that It varies predictably with very smal I stimulus changes (Davis & Heninger, 1972). Lesion and electrical stimulation studies in the rat have shown that the acoustic startle response Is mediated by a relatively simple neural circuit (Davis, Gendelman, Tischler, & Gendelman, 1982). Auditory input ascends via the auditory nerve and the ventral cochlear nucleus to the nuclei of the lateral lemnlscus, which connect to the nucleus reticular is pontis caudal Is, and this is the beginning of the motor output component of the startle pathway. Startle is a brainstem reflex that can be found in decerebrate animals (Davis & Gendelman, 1977; Fox, 1979), and in anencephal ic infants (W. K. Berg, unpubl !shed data). Startle has been measured in adult and infant animals and humans, using a variety of techniques. In animal research, the two nost common measures are electromyographic (EMG) activity of the flexor muscles of the leg, and the whole-body jerk, as measured by an accelerometer In a rigidly suspended cage. In humans the eyebl Ink component is assessed, either by measuring the movement of the eyelid with a potentiometer, or by measuring the EMG activity of the orbicularls ocul i, the muscle responsible for the eyebl Ink. The advantage of using the eyebl Ink

PAGE 15

8 component of startle in the human is that it has a lower threshold than either flexor EMG or whole-body jerk, so minimal ly-startl Ing stlmul I can be used. The EMG measure may be preferred to I id movement, since EMG activity can be found In the absence of I id movement, whether the eyes are open or closed, making EMG both more sensitive and more versatl le than I id movement. Startle responses can be assessed by the appl !cation of two different paradigms. First, the direct effects of stlmul I on the startle response can be studied, by varying the eliciting stimuli and measuring the resultant changes in the startle response. Second, the modification of the startle response by a nonstartl ing stimulus (cal led a prepulse), occurring before the eliciting stimulus, can be measured. The first technique Is aimed at a direct assessment of the startle system, while the second technique uses the startle stimulus as a probe to Investigate processing of the prepulse. The degree to which a stimulus wil I el iclt startle is affected by several factors. Increasing stimulus rise time results In higher startle thresholds, In both rats (Fleshier, 1965) and human adults (Berg, 1973). Increasing rise time also results In smaller bl inks for suprathreshold stimuli in human adults (Blumenthal & Berg, 1984) and infants (Blumenthal & Berg, 1982). Increasing stimulus Intensity yields larger responses in adults (Blumenthal & Berg, 1984) and In rats (Marsh, Hoffman, & Stitt, 1973). Increasing stimulus duration up to 8 ms, but not beyond, yields larger responses in rats (Marsh, Hoffman, & Stitt, 1973), and this critical duration decreases as stimulus Intensity Increases. In human adults, the critical duration Is longer than it is in rats. Berg (1973) increased stimulus duration up to 32

PAGE 16

9 ms, and found a continuing increase in response amp I itude. Yamada (1983) found that response amp I ltude increased as stimulus duration Increased to 30-50 ms, but not beyond. This i I lustrates a critical duration for human adults which is considerably higher than that for rats. The critical duration for Infants has not previously been reported. In rats, Increasing the interval between two brief (1 ms) pulses to 3 ms (onset-to-onset) results in larger responses, with a return to the level of a single 1 ms pulse when lnterpulse interval is increased to 6 ms (Marsh, Hoffman, & Stitt, 1973). The increase In responding as interpulse interval Is increased to 3 ms may be due to a reduction of neural refractoriness and more complete peripheral recovery. At interpulse intervals greater than 6 ms, the second pulse appears to have no effect, II lustratlng a critical Interval for summation of transients of 6 ms in rats. The startle modification technique uses the startle stimulus as a probe, and can be used to assess stimulus processing In several modal itles. This method Involves the presentation of the stimulus to be assessed, the prepulse, some brief time before the startle stimulus. Depending on the stimulus parameters of the prepulse, the startle response can be Inhibited or facilitated in several ways. The amount of modification Is who I ly dependent upon the parameters of the prepulse, with the startle-el !citing stimulus having no effect on the modification (Stitt, Hoffman, & Marsh, 1976). The time between onset of the prepulse and onset of the startle pulse ls cal led lead time, and this Is an important factor In determining modification effects. At lead times of 10 ms or less in the

PAGE 17

10 rat, or 30 ms or less In the human, a prepulse above detection threshold wil I cause a decrease In the latency of the subsequent startle response (Hoffman & Ison, 1980). At lead times of 30-240 ms in both rats and humans, startle amp I ltude is reduced by the prepulse (Hoffman & Ison, 1980). The optimal lead time, where the amount of Inhibition of amplitude is maximal, is about 120 ms In adults. If lead time Is equal, prepulses which remain on throughout the lead tine are no more effective than those which have a duration of 20 ms (Graham & Murray, 1977). However, as the duration of transient prepulses increases, more amp I ltude inhibition is found. This may be because the onset and offset of the transient prepulse become more independent as they are separated in time, so their combined effect Is Increased * (Harbin & Berg, 1984). In infants, the transient-sensitive system Is bel leved to be immature (Graham, Strock, & Zeigler, 1981). For maximal Inhibition of startle amp I ltude by a transient prepulse in 6-to-9-week-olds, lead time must be 225 ms, as compared with 120 ms In adults. More intense prepulses are required with these Infants, and the prepulses must be * If a brief, transient stimulus occurs at a very long lead time, It wil I have no effect, unless attention Is directed to the prepulse (as indicated by heart rate deceleration) (Bohl in, Graham, Silverstein, & Hackley, 1981). If attention Is directed towards the sensory modal lty in which the startle stimulus occurs, the startle response wll I be facil ltated; if attention Is directed away from the startled modal lty, Inhibition occurs. The attentional effect can combine with the effect of a sustained prepulse, resulting in more facll itatlon than to either effect alone (Bohl In, Graham, Silverstein, & Hackley, 1981). The reason for this attentional effect may be that heart deceleration, or orienting, may cause an enhancement of stimulus Input so that the effectiveness of a startle stimulus occurring during orienting to a prepulse Is increased (Lacey & Lacey, 1974).

PAGE 18

11 slightly longer (25 30 ms in Infants, 20 ms In adults) In order to be effective (Zeigler, Strock, & Graham, 1979). Even at optimal lead times, Inhibition Is less marked in Infants than in adults (Graham, Strock, & Zeigler, 1981). Since startle inhibition is argued to be a function of transient system activity, these findings suggest that the temporal summation of transients ls Immature In infants. At very long lead times (2-4 sec) a prepulse which stays on throughout the lead time (sustained prepulse) can cause an Increase In startle amp I ltude in adults and infants, but optimal lead time is longer for infants (4 sec) than for adults (2 sec). However, the degree of facil ltation Is similar at the two ages, supporting the bef ief that the sustained processing system ls relatively mature In Infancy (Graham, Strock, & Zeigler, 1981). In rats, startle Inhibition and facil ltatlon have been Investigated at the neurochemlcal level and appear to be due to the activation of different receptor sites. The receptors responsible for facil ltatfon of startle appear to mature earlier than those responsible for inhibition (Gal lager, Kehne, Wakeman, & Davis, 1983). If an analogous arrangement exists In the human, developmental differences in startle modification may be expected. Immaturity of the transient system has been proposed, based on evidence using the startle modification paradigm (Graham, Strock, Zeigler, 1981). A more direct assessment of the transient system may be possible by using the startle el !citation paradigm, Investigating stimulus effects on the startle response directly, rather than Its modification by a prepulse (Dykman & Ison, 1979). Also, If transient system immaturity ls believed to exist, It may be most apparent In

PAGE 19

12 newborns, so this age group, rather than older infants, should be studied. In the present study, the temporal summation of transient and sustained stimul I was assessed In human adults and newborns, using the startle el icltatlon technique. The stimuli used were either single stimuli, varying in duration (3-100 ms), or pairs of brief (3 ms) stlmul I, with onset-to-onset Intervals analogous to the duration of the single stimuli. Since peripheral recovery requires more than 3 ms (Green, 1973), the offset of very brief pulses may have a minimal effect, so the paired stlmul I were treated as two transients rather than four. These stimul I were used to assess transient system activity directly. The single stlmul I longer than 3 ms Included transients (at onset and offset) and a steady-state portion (between onset and offset) so that responding to these stlmul I could be contributed to by both transient and sustained system activity. The contribution of the sustained system could be assessed In three ways, al I of which Involve the assumption that the onset of single stimuli has an effect equivalent to that of the first pulse In a pair of brief pulses, that is, transient system activity. The assessment of sustained system activity differs in how the stimulus components fol lowing onset (i.e. steady-state portion and offset) affect responding. Stimulus offset may be assumed to have no effect on the startle response, if startle is a function of a sufficient Increase In stimulus Intensity occurring within a sufficiently brief time, as Graham (1975) states. Therefore, subtracting the response to a 3 ms pulse from that to a longer pulse (subtraction of onset transient) would yield the

PAGE 20

13 effect of the steady-state portion of the stimulus. However, the influence of stimulus offset on startle has irrelevant, so Ignoring this possible effect not been shown to be may result In an overestimation of sustained system activity (since some of the response remaining may be due to transient system activity). A second way to infer sustained system activity is to subtract the response to a pair of transients from the response to an analogous single stimulus. This assumes that offset Is as powerful a transient as onset, and that startle Is determined by sufficient stimulus change in either direction. With this method, sustained system activity wll I be underestimated if offset is not as effective a transient as onset (since transient system activity wl I I be overestimated). Since the effect of stimulus offset has not been quantified, it may be useful to consider the situation where the effect of the offset transient Is greater than zero but less than that of the onset (or the 3 ms pulse). If a single stimulus is no more effective than an analogous pair of stlmul i, It is assumed that the steady-state portion and offset of the single stimulus are no more effective than the second pulse In a pair. The degree to which single stimuli are more effective than paired stlmul i Indicates sustained system activity, though the ful I effect cannot be ascertained unless offset transient effects can be determined. However, If stimulus duration Increases and responding does not change, sustained system activity may have a minimal contribution. This does not necessarily mean that sustained activity Is not occurring, it just means that this activity is not contributing much to the response. When both transient system activity and sustained system activity are present, their contributions to the response may be

PAGE 21

14 unequal. If the activation of the more efficient system is decreased, the activity of the other system may become more apparent The only case in which the contribution of sustained system activity would be ambiguous is that for paired stlmul i which produce larger responses than single stimuli. The lower responding to single stlmul i could be due to either reduced sustained system activity or to reduced transient system activity at onset. To assess the activity of the sustained system more directly, research using stimul I with minimal transients (long rise and fal I times) is needed. The effects of stimulus onset alone could be examined with fast-rising, slow-fal I Ing stlmul i, so that no offset transient occurs. Stimulus offset effects could be isolated with slow-rising, fast-fal I ing stimuli, to minimize the onset transient. These avenues of research are suggested to explore the sustained processing system more fu 11 y.

PAGE 22

Subjects CHAPTER I I EXPERIMENT 1 Method Subjects were 16 adults, 6 males and 10 females, with a mean age of 20 years (range=18 to 27 years), selected from a university undergraduate subject pool, who reported no history of hearing loss and no use of medications on the day of testing. Stimul I Broadband noise bursts (20Hz-20kHz), at 95 dB(A), with a rise-fal I time of less than 0.1 ms were used. Seven of the stimul I varied In duration (3, 20, 35, 50, 65, 80, and 100 ms), and six others consisted of pairs of 3 ms pulses at intervals analogous to the durations used (20, 35, 50, 65, 80, and 100 ms, measured from onset to onset). lntertrial interval ranged from 25-35 sec, averaging 30 sec. Apparatus Stlmul i were produced by a Grason-Stadler 455C noise generator gated through an Icon ix electronic switch and a Sansul AU517 amp I ifler, and presented through a JBL Decade 26 loudspeaker, located approximately 1.5 m in front of the subjects. Stimulus intensity was calibrated using a General Radio 1551C sound level meter and monitored with a Hewlett-Packard 400E AC voltmeter. 15

PAGE 23

16 Perl-orbital electromyographic (EMG) responses were collected using miniature Beckman blopotentlal electrodes (Ag/AgCI) filled with Synapse conducting paste. The EMG signal was amp I ified by a Coulbourn Hi-Gain Bloampl ifler/Coupler, with filters passing frequencies of 90250 Hz, and a gain of 63,000. The signal was then integrated with a Coulbourn Contour Fol lowing Integrator at a time constant of 80 ms * (integrated EMG). The Integrated EMG was then recorded on a Beckman R411 polygraph, as wel I as being dig Ital ly sampled (10 bit accuracy) by a PDP-8 computer every mil I !second for 250 ms after stimulus onset. Procedure The experimenter explained the procedure and the subject was asked to read and sign an informed consent statement, and to fil I out a background questionnaire. The experimenter then cleaned the area just below the left eye with a cotton swab soaked in alcohol, and two electrodes were attached, one below the center of the eye, and the other Immediately temporal to the first, as close to the orbital ridge as possible without impairing eye movement. The subject was then seated In the testing room and asked to move as I ittle as possible, especially the eyes and head. Data were used only from trials on which an experimenter, watching the subject on closed circuit television, judged the eyes to be open. If a stimulus was presented when the eyes were closed, or during movement of the head or eyes, the trial was rejected * The integrator time constant was determined by cal ibratlon, not by the setting on the Integrator itself. It was found that the time constant settings of two integrators of the same model varied widely, so the only rel I able way to determine the actual time constant was to cal lbrate each separately.

PAGE 24

17 by an experimenter bl ind to stimulus conditions. An average of 7.5 percent of the trials were rejected for each subject, and rejection was not more I ikely for one stimulus condition than for any other. A 13x13 Latin square was used to determine stimulus order, with the row of entry into the square being randomly determined. A session was terminated when six potentially scorable responses In each stimulus condition were obtained. Q.a.:ta Analysis Response amp I itude was measured as the difference between response onset and peak. Response onset was judged only during the window between 20 and 100 ms after stimulus onset. Initiation of a response was judged on the basis of a monotonic increase in EMG that continued for at least 20 ms. Onset latency was then judged as the point where this increase exceeded the random variability in the 20 ms prior to this increase. The peak of the response was the first point fol lowing onset where the slope of the Integrated EMG signal equal I ed or passed through O, when this was fol lowed by at I east 10 ms with no further slope reversa I. Response latency was measured as the time from sti mu I us onset to response onset. Response probability was measured as the degree to which responses actually occurred when subject variables were optimal (i.e. eyes open, no head movement). If the integrated EMG signal during the response window did not deviate beyond the random noise present prior to the onset window, a failure to respond was recorded. Due to the considerable sensitivity of the computer sampling and amp I lfier, the degree to which very smal I responses were not detected was minimal.

PAGE 25

18 The response parameter most often reported by researchers In this area Is average magnitude, which is calculated from trials on which a response could have been recorded, whether a response actually occurred or not (e.g. Graham & Murray, 1977). Magnitude changes can be due to changes In either response probabll ity or amp I itude (Prokasy & Ebel, 1967), and these two may be partially independent measures of startle responding (Blumenthal & Berg, in press). Therefore, the present study examined response amp I itude and probability separately. For each subject, response amp I ltude and latency were averaged across trials for each condition. ANOVAs Including stimulus type (single or paired) and time (duration or interval in msec) as within-subject variables were conducted. To balance the analyses, data from the 3 ms single stimulus condition were duplicated for use in paired stimulus comparisons, as a substitute for a pair of pulses at an Interval of Oms. Results Amp I itude The effect of increasing stimulus duration or interpulse Interval on response amplitude differed, as ii lustrated by a slgnif icant stimulus type by time interaction, E (6,90)=8.88, ~<.001 (see Figure 1). Increasing stimulus duration to 50 ms, but not beyond, resulted in larger responses, as shown by significant I inear, E (1,15)=18.23, ~<.001, and quadratic, E (1,15)=15.37, ~<.001, trends. Responses were larger at 20 than at 3 ms duration, i (15)=4.13, ~<.001, and larger at 50 than at 35 ms duration, i (15)=2.64, ~<.025. No other pairwise comparisons reached significance, indicating that increasing duration

PAGE 26

19 55 50 en 45 SINGLE C ::, ... 0 ... 40 .0 ... 0 w 0 35 :::> ._. _j 30
PAGE 27

20 above 50 ms had no significant effect. Increasing interpulse Interval to 35 ms, but not beyond, resulted In larger responses, yieldlng a significant quadratic trend, E (1,15)=5.61, ~<.05, and a marginally significant cubic trend, E (1,15)=4.46, ~<.052. Responses were larger for a pair of pulses at 20 ms interval than for a single 3 ms pulse, 1 (15)=2.20, ~<.05. Increasing Interval from 3 to 35 ms resulted In larger responses, 1 (15)=2.85, ~<.025, but responses were not signlflcantly larger at 35 than at 20 ms interval. At intervals of 50 ms or more, responses were not slgnificantly larger than for a single 3 ms pulse. Up to 35 ms, a single stimulus was not significantly more effective than a pair of transients, showing that, even with much less total energy, a second transient was as effective as the sustained portion and the offset of a slngle stlmulus. This Indicates an asymmetrical Input of transient and sustained stlmul i to the startle system, with transient input being more effective than sustained Input, up to 35 ms. Above 35 ms, the second transient had I ittle effect, whereas the sustained portion of the stimulus Increased in effectiveness, reaching asymptote at 50 ms. Probability For response probabll lty, the effect of increasing stimulus duration or lnterpulse Interval also differed, as II lustrated by a significant stimulus type by time Interaction, E (6,90)=5.34, ~<.001 (see Figure 2). A I lnear effect of duration appeared, E (1,15)=13.29, ~<.01, and this was due mainly to increasing duration from 3 to 20 ms, 1 (15)=2.22, ~<.05. In fact, if the 3 ms data were excluded from the analysis, no significant duration effect was found, E (1,15)=2.12. For

PAGE 28

100 90 C Q) 0 Q) 0.. 80 >t_J 70 CD 0 0::: a.. w 60 en z 0 a.. 50 0::: 1, 3 20 21 35 50 ' ' 65 80 SINGLE PAIRED 100 DURATION/ INTERVAL ( ms) Figure 2. Adult response probability as a function of time for single and paired stimuli.

PAGE 29

22 paired stlmul i, a quadratic trend appeared, E (1,15)=6.16, ~<.025. Increasing interpulse Interval to 50 ms had no further effect. At longer Intervals, the contribution of the second pulse disappeared, with probability at intervals of 65, 80, and 100 ms not significantly different from that for a single 3 ms pulse. Up to 50 ms, the difference In probability between single and paired stimuli was not significant, ii lustrating a dominant transient contribution. When the second transient was no longer effective, the sustained portion of the stimulus maintained a high response probabil lty. However, the Interpretation of these probabii lty data may be influenced by a cell Ing effect for single stlmul I. Latency Response latency was affected by increasing stimulus duration or lnterpulse interval, E (6,90)=2.22, ~<.05, and stimulus type had a significant effect, E (1,15)=10.40, ~<.01, but the stimulus type by time Interaction was not significant (see Figure 3). For single stimuli, Increasing duration had an effect on latency, E (6,90)=4.43, ~<.001, but this was due to the increase from 3 to 20 ms only, since, if the 3 ms data were not included in the analysis, no duration effect was found ([ (5,75)=0.64). For paired stimuli, increasing lnterpulse interval had no effect on response latency. Discuss loo The findings of Experiment show that temporal summation of transient acoustic Input dominates that of sustained Input, In adults. Transients occurring within a window of 35-50 ms summate, and sustained

PAGE 30

23 62 1', I \ I \ I .................. .......... PAIRED 60 \ \ I I \ \ I 58 en E I >(.) z 56 w ....J w SINGLE Cf) 54 z 0 a.. Cf) w a::: 52 50 3 20 35 50 65 80 100 DURATION/ INTERVAL (ms) Figure 3. Adult response latenc y as a function of time for single and paired stimuli.

PAGE 31

24 summation adds I ittle which cannot be explained by transient summation In this range. Beyond this window, transient summation does not occur, and the contribution of sustained summation becomes apparent. Also, these differential summation effects are reflected by startle response ampl ltude, probabll ity, and latency in different ways. For response amp I ltude the summation window for transients is 3550 ms, and the time between transients has a slight, but not significant, effect on the response. As interpulse Interval within the window increases, a slight Increase In response amp I ltude may be explained by forward masking, which suggests that the first pulse decreases the effectiveness of the second. As the interpulse Interval Increases the amount of masking decreases (Raab, 1961), so that responding at the 20 and 35 ms intervals may differ due to a reduction in responding at the 20 ms interval, but not at the 35 ms Interval (or, at least, not as great a reduction). If forward masking is present, its Influence is not great, possibly due to the brief duration of the first pulse (Zwislocki, 1978). Also, this marginal masking effect was found for response amplitude, but not for response probability. For response probability, the summation window extends to 50 ms, and the effect of the second transient Is not Influenced by where in this window It occurs, above 20 ms. For response latency, the second transient had no apparent effect at any of the intervals used. It may be the case that the temporal summation window for transients Is narrower than 20 ms, so that the second pulse always occurred outside the window, and this prediction may be supported by future research. For sustained summation, 20 ms Is adequate, but the window may be less than this.

PAGE 32

25 After the transient summation window has been exceeded, sustained summation maintaining continues elevated to occur, responding either after the increasing effect responding or of the second transient has disappeared. This peaks at 50 ms for response ampl Jtude, and at 20 ms for response latency. For response probabil lty, the peak of the sustained summation function is difficult to specify due to the presence of a celling effect at high probabilities, but this peak could not exceed the peak response latency (approximately 65 ms). Also, the extent to which the second pulse in a pair el !cited a response when the first pulse failed to do so was minimal (less than 10 Instances for al I subjects combined). This means that the second pulse alone Is not effective in el !citing responding. The temporal summation of transient stlmul I peaked at 35 ms for response amp I ltude and 50 ms for response probabll ity. Within these ranges, the contribution of the sustained portion of a stimulus was no greater than that of the second brief transient. Beyond these ranges, the second stimulus in a pulse pair had I lttle or no effect, but responding was maintained by the sustained portion of a single stimulus. A second experiment was conducted, to investigate these temporal summation properties In greater detail.

PAGE 33

The data of Experiment CHAPTER 111 EXPERIMENT 2 Introduction suggest that the temporal summation window for transient stlmul i Is 35-50 ms for response amp I ltude and probabll ity, and less than 20 ms for response latency. To more accurately assess responding in the critical range of 35-50 ms, a second experiment was conducted, In which stimulus duration and interpulse intervals In the 30-55 ms range were varied in 5 ms steps. Method Subjects Subjects were 12 university students, 6 males and 6 females, with an average age of 22 years (range=18 to 29), selected in the same way as In Experiment 1. Stiroul I, Apparatus, Procedure, and Data Analysis These were identical to those used In Experiment 1, except that stimulus durations were 3, 30, 35, 40, 45, 50, and 55 ms, with analogous lnterpulse Intervals for paired (3 ms) stlmul I. Results Amp I itude Response amp I itude was affected by Increasing stimulus duration or lnterstimulus interval, E (6,66)= 3.70, <.005, and stimulus type had a significant effect, E (1,11)= 29.71, <.001, but the two variables 26

PAGE 34

27 did not interact (see Figure 4). For single stimul I, a significant I inear trend for duration was found, E (1,11)= 9.70, <.01. A single 30 ms stimulus was more effective than a 3 ms stimulus, I (11)=2.86, <.025, but further Increases In stimulus duration were not significantly more effective than a 30 ms stimulus. For paired stlmul I, a significant quadratic time trend was found, E (1,11)= 6.81, <.025. Pairs of pulses at Intervals of 35, 40, and 45 ms were sl lghtly, but not significantly, more effective than a single 3 ms stimulus ( I (11)= 2.09, <.06; i (11)= 1.98, <.08; i (11)= 2.02, <.07, for 35, 40, and 45 ms, respectively). At 50 and 55 ms intervals, pairs of pulses were clearly no more effective than single 3 ms stimuli. Single stlmul I were significantly more effective than paired stimul I at 50 ms, I (11)= 3.44, <.01, but not below. Probabll ity The effect of Increasing stimulus duration or interslmulus interval differed for response probability, as II lustrated by a significant stimulus type by time Interaction, E (6,66)= 2.27, <.05 (see Figure 5). For single stlmul I, a I lnear trend for duration was found, E (1,11)= 7.53, <.025. A 30-ms duration stimulus was more effective than a 3 ms stimulus, i (11)= 2.40, <.05, but no significant difference was found between a single 3 ms pulse and pairs of pulses at al I other Intervals tested. For paired stlmul I, a significant quadratic interval trend was found, E (1,11)= 5.64, <.05. A pair of pulses 30 ms apart were more effective than a single 3 ms stimulus, i (11)= 2.52, <.05, but further increases in interstlmulus interval had no significant effect. A single stimulus was more

PAGE 35

w 0 => t_J (/) a.. C :::, <( w .... 0 (/) .... z .Q 0 .... a.. 0 (/) w a: 45 40 35 30 25 20 3 .,,,.,.,... 28 I I \ SINGLE \ \ / PAIRED \I 30 3540 45 50 55 DURATION/ INTERVAL ( ms) Figure 4. Adult response amplitude as a function of time for single and paired stimuli.

PAGE 36

al al 0 a: -= Ii. C Q) LLJ 0 en z a. o_ Ii. en LLJ a: 80 70 60 50 3 29 ......... / '---" "PAIRED 20 25 303540 45 DURATION/ INTERVAL ( ms) Figure 5. Adult response probability as a function of time for single and paired stimuli.

PAGE 37

30 effective than an analogous pair of stlmul i at 40, 50, and 55 ms (I (11)= 2.32, 2.28, and 2.54, respectively,~ <.05). Latency A significant stimulus type by time interaction was found for response latency, E (6,66)= 2.54, <.05 (see Figure 6). For single stimul I, the overal I time effect was not significant, but latency was lower for 40 ms stlmul I than for 3 ms stimuli, I (11)= 3.39, <.01. For paired stlmul I, a I lnear effect of lnterstlmulus Interval was found, E (1,11)= 14.65, <.005. Increasing interval from 30 to 50 ms resulted In a significant increase in response latency, I (11)= 2.75, <.025, but no other comparisons of paired stimuli were significant. Latencies for single stimul I were significantly lower than those for paired stlmul i, at each Interval tested. Discussion For response amp I itude, Increasing stimulus duration to 30-55 ms resulted In larger responses, as II lustrated by a significant I inear trend for duration. Increasing lnterstimulus interval for paired stimul I resulted In larger responses up to 40 ms, then a reduction In response ampl ltude, as shown by the significant quadratic function for paired stimuli. Single stimuli were not significantly more effective than paired stimuli below 50 ms. In Experiment 2, the two functions diverged between 35 and 50 ms, just as they did In Experiment 1. The fact that the response amp I ltude functions are so similar in the two experiments suggests that these findings are rel I able. Up to 35 ms, the transient aspects of the stimulus seem almost as effective as the

PAGE 38

31 68 I\ 66 I \ I \ f/) PAIRED E 64 / >/ u /"""'v z 62 l1J / t/ <{ / _J 60 l1J Cl) z 58 0 SINGLE Q.. Cl) 56 l1J 0:: 54 1 3 30 35 4045 50 55 DURATION/ INTERVAL (ms) Figure 6. Adult response latency as a function of time for single and paired stimuli.

PAGE 39

32 sustained aspects in determining response amplitude. In Experiment 1, It was shown that temporal summation for response probabll lty may peak below 20 ms, and this was supported In Experiment 2. Increasing stimulus duration from 3 to 30 ms resulted In Increased response probability, with further increases to 55 ms having no greater effect. A pair of stimuli 30 ms apart were more effective than a single 3 ms pulse but further Interval Increases were not effective. Paired pulses were as effective as single pulses up to 45 ms, above which the effect of the second pulse In the pair tended to decrease. This para! lels the finding in Experiment 1 that transient summation Is as effective as sustained summation up to 50 ms. The response latency data of Experiment 2 are also similar to those of Experiment 1, but not as clearly as were the amp I itude and probabll ity data. Latency was lower for single than for paired stimuli, In al I cases. However, where the latency effect for single stlmul I peaked at 20 ms In Experiment 1, it peaked at 40 ms In Experiment 2. Also, where no effect of latency was found for paired stimuli In Experiment 1, longer latency was found at 50 than at 30 ms Intervals in Experiment 2. The reason for these latency differences In the two experiments is not clear, but these data suggest that latency may be a less reliable measure of startle activity than either response amp I itude or probability. Experiments and 2 show that the temporal summation of acoustic stlmul i occurs in different ways for transient and sustained stimul I, and that this differential summation is reflected In different ways by startle ampl ltude, probability, and latency. For response amplitude, transient summation dominates for 40 ms, and sustained summation adds

PAGE 40

33 I ittle during this time. Beyond the transient summation window, sustained summation continues to increase response amp I ltude for a brief time (<20 ms). For response probability, summation of transients peaks at less than 20 ms, and continues to 50 ms, with minimal contribution by sustained summation in this period. Above 50 ms, sustained summation maintains elevated probability while the effect of a second transient disappears. The probability and ampl ltude functions are similar in that the temporal summation window for transients is not more than 50 ms In both cases, but this summation seems to peak earl ler for probability than for amplitude. Response latency appears to reflect summation properties similar to those shown by probability, but In a less reliable fashion. Experiments and 2 ii lustrate the fact that the startle el icitatlon paradigm provides an effective tool for comparing transient and sustained system activity. Therefore, it would be of Interest to apply this paradigm to infant subjects, since Graham, Anthony, & Zeigler (1983) hypothesize that the transient system ls relatively immature when compared to the sustained system in infants. Experiment 3 measured startle amp I ltude, probability, and latency to stlmul I analogous to those used in Experiments t and 2. If transient immaturity exists, these measures should show differences between Infants and adults.

PAGE 41

Subjects CHAPTER IV EXPERIMENT 3 Method Al I subjects were apparently healthy newborns, with no significant problems or comp I ications, and were the result of spontaneous vaginal delivery at Shands Hospital in Gainesville, Florida. Data from 4 of 24 subjects tested were not used, due to a failure to respond to at least two trials at each stimulus condition. For the 20 remaining Infants, average age was 23 hours (range=7-36 h), average gestational age by maturity rating was 39.75 weeks (range=37-42 wk), average 5 minute Apgar (1953) score was 8.9 (range=8-10), and average birthwelght was 3350 gm (range=2640-4220 gm). Testing began an average of 1 hour 10 minutes after the last feeding (range= 30 min3 h 30 min). Local anaesthetic was administered to 11 mothers during delivery, with no anaesthetic In the remaining 9 cases. Apparatus Stimuli were produced by a Grason-Stadler 455C noise generator, gated through a Coulbourn electronic switch and a Dynaco Stereo 120 amp I lfier, and presented through a JBL Decade 26 loudspeaker, located 1 m In front of the subjects. Sound levels were checked before each session, using a General Radio 1551C sound level meter and a Simpson 260 voltmeter. The EMG signal was amp I if led and integrated In the same way as in Experiment 1, except tor some variation In Bioampl ifler gain 34

PAGE 42

35 settings. This signal was sampled by an Apple l I Plus computer (12 bit accuracy) every millisecond for 250 ms and corrected for gain variations. EKG was recorded on a Teac A-2300SX tape recorder, and on a Narco DMP-4B Physlograph. Respiration was recorded using a Parks Electronics strain gauge and a Parks Electronics Model 270 plethysmograph, and this signal was also displayed on the Narco Physlograph. Stimu Ii Stimuli were broadband noise bursts (20Hz-20 KHz) at 100 dB(A) with fast(< 0.1 ms) rise and fal I times. Pilot data showed that 95 dB(A) stimuli (the intensity used In Experiments and 2) did not rel I ably elicit responding at 3 ms duration. Five of the stlmul I varied In duration (3, 20, 35, 50, and 100 ms), and the other four were pairs of 3 ms pulses at onset-to-onset Intervals of 20, 35, 50, and 100 ms. lnterstlmulus interval ranged from 20-60 ms. Procedure Each subject was placed supine In a crib In the testing chamber, and the respiration strain gauge was attached laterally across the abdomen, just above the navel. EKG electrodes were attached In a modified Lead I l configuration. The area Just below the left eye was then cleaned with a cotton swab soaked in alcohol, and the EMG electrodes were attached, one just below the center of the eye and the other immediately temporal to the first, as close to the orbital ridge as possible. Stimulus presentation was control led by an experimenter, sitting beside the loudspeaker, who also recorded the state of the Infant on each trial, using the rating scale of Berg, Berg, and Graham

PAGE 43

36 (1971). These behavioral ratings, along with respiration and EKG measures, were used to determine the state of the infant, and these three measures have been shown to be adequate and reliable Indicators of sleep state in young infants (Anders, Emde, & Parmelee, 1971). Only trials during which the Infant was judged to be in quiet sleep were Included In the analysis. Stlmul i were not presented during movement of the head, eyes, or arms, or when the subject's eyes were open. A session was terminated when 9 potentially scorable responses at each stimulus condition were obtained. If the Infant changed state the session was halted, and the infant was al lowed to return to quiet sleep. If the subject did not reach quiet sleep within an hour, the truncated session was terminated. Stimulus order was determined by a 9x9 Latin square, with the row of entry into the square being randomly determined. Response amp I itude, probabll ity, and latency were measured in the same way as in Experiment 1. Results Amp I itude A significant effect of stimulus type was found, E (1,19)=5.13, ~<.05, with single stlmul I el !citing larger responses than paired stimuli (see Figure 7). A significant time effect was also found, E (4,76)=6.46, ~<.001, but stimulus type and time did not Interact. For single stlmul i, Increasing duration from 3 to 20 ms yielded larger responses, i (19)= 2.09, ~<.05, and a further increase from 20 to 35 ms resulted In still larger responses, i (19)=4.07, ~<.001. Increasing stimulus duration above 35 ms had no further effect. For paired stimul I, increasing interpulse interval from 20 to 50 ms resulted In larger responses, i (19)=2.12, ~<.05. Responding to a pair of pulses at

PAGE 44

37 100 (/) 90 -+C ::::, >.. SINGLE '0 '-+!', .0 '0 80 ...__, I " w " 0 I '-..__ :::J II "-. PAIRED _J 0.. I >70 4 I w I U) z I 0 0.. I U) 60 w I 0::: I 50 3 20 35 50 100 DURATION/ INTERVAL ( ms) Figure 7. Infant response amplitude as a function of time for sin g le and paired stimuli.

PAGE 45

38 an interval of 100 ms was not significantly different from that to a single pulse at 3 ms duration, suggesting that the second pulse In the pair did not contribute to the response. A single stimulus 20 ms long was not significantly more effective than a pair of pulses 20 ms apart. At 35 ms, single stlmul I were more effective than pairs of pulses, (19)=2.47, ~<.025, but this advantage was not found at 50 or 100 ms. These findings suggest that a pair of pulses can be nearly as effective as an analogous single stimulus, up to 50 ms. The transient contribution seemed to be dominant In determining responding, with the sustained portion of the stimulus adding I lttle. Probabll lty For response probabil lty, an Interaction between stimulus type and time was found, E (4,76)=20.67, ~<.001 (see Figure 8). For single stimuli, a quadratic duration trend appeared, E (1,19)=24.28, ~<.001. Response probability increased as stimulus duration was Increased from 3 to 20 ms, i (19)=2.76, ~<.025, and from 20 to 35 ms, (19)=4.41, ~<.001, with duration Increases above 35 ms having I ittle effect. For paired stimul I, a quadratic interval effect appeared, E (1,19)= 19.79, ~<.001. Probabil lty was not affected by Increasing lnterpulse interval to 50 ms. At 100 ms Interval, probabll lty returned to the level found with a single 3 ms pulse, suggesting that the second pulse at 100 ms did not contribute to responding. A pair of pulses at 20 ms Interval was more effective than a single 3 ms pulse, i (19)=2.82, ~<.01, and no less effective than a 20 ms long stimulus. Single stlmul I were more effective than paired pulses

PAGE 46

100 -90 C Cl) u Cl) 80 >._ _J m 70 <{ m 0 0::: a.. w Cf) z 0 60 a. 50 Cf) w 0::: 3 39 ________ SINGLE 20 35 50 100 DURATION/ INTERVAL ( ms) Figure 8. Infant response prob a bility as a function of time for single and paired stimuli.

PAGE 47

40 at 35, 50, and 100 ms, indicating a slgnif icant contribution of sustained stimulus attributes above 20 ms. Latency A significant effect of stimulus type on response latency was found, E (1,19)=4.53, ~<.05, with latency being slightly longer for paired than for single stlmul I (except at 100 ms) (see Figure 9). No significant time effect or stimulus type by time interaction was found. Pairwise comparisons, looking at time comparisons for eoch stimulus type, and stimulus type comparisons for each time, showed that only the 50 ms single and paired stlmul i resulted in different response latency, i (19)=2.42, ~<.05. These data are difficult to Interpret, due to the high latency variability, an effect also reported by Zeigler (1978) In 6-to-9-week-old Infants. Discussion The present study suggests that the temporal summation of auditory stimul I in human newborns fol lows a different time course for transient and sustained stlmul I. Also, startle amp I itude, probabll ity, and latency do not reflect this differential summation in the same way. Response amp I ltude increases if a second transient is added up to 50 ms after the first. In general, the sustained portion of the stimulus does not seem to contribute much that cannot be accounted for by the transient components of the single stimuli. This suggests that the window for temporal summation of transients Is 50 ms for response amp I ltude, and that transient summation dominates sustained summation in this range.

PAGE 48

41 60 ;\ 58 I \ I \ I \ en E 56 I \ >I \ u z --\ w 54 ---\ PAIRED _J \ w \ (/) 52 z \ 0 a.. \ (/) w a:: 50 SINGLE \; 1, 3 20 35 50 100 DURATION/ INTERVAL {ms) Figure 9. Infant response latency as a function of time for single and paired stimuli.

PAGE 49

4 2 The fact that response amp I itude increases as interpulse interval increases to 50 ms suggests that forward masking may be present, with the first pulse in the pair decreasing the response to the second pulse. As interpulse interval increases, masking decreases (Raab, 1961), al lowing the second pulse to contribute more to the response. The findings for response probability are quite different. Adding a second transient 20-50 ms after the first results in Increased I ikl ihood of responding, and the effect of the second transient ls not influenced by where It occurs within this window. Sustained summation becomes Important above 20 ms, resulting In an increase in probabil lty which asymptotes at 55 ms. This suggests that the window for temporal summation of transients Is 20-50 ms, and that sustained summation above 20 ms can add to this, with this sustained contribution reaching Its maximum within 35-50 ms of stimulus onset. The effect of the second transient is less pronounced for probabll ity than for response amp I itude, so transient and sustained system activity coexist, and both contribute significantly to responding In infants.

PAGE 50

CHAPTER V GENERAL DISCUSSION The startle response is Influenced by activity In both the transient and sustained systems, a finding contrary to the hypothesis of Dykman and Ison (1979). These two systems are at least partially independent, with the startle response being the net result of activity in both systems. These experiments show that manipulation of the startle-el icltlng stlmul i can demonstrate Independent Influences of transient and sustained system activity. However, the relative contribution to the startle response of these two systems may not be equal, since these results suggest that the transient system Is more efficient than the sustained system. As one system becomes dominant over the other, their relative contributions to the response wll I change. To find the relative contribution of transient and sustained activity, it is first necessary to calculate the effect of a second 3 ms pulse in a pair if no transient activity were to occur for that pulse. This is accompl lshed by calculating the slope of the single stimulus function (in units per ms), and adding 3 times this value to the response to a single 3 ms stimulus. If responding to two pulses is simply additive, then 6 ms of stimulation should be equivalent to two 3 ms stimuli. The extent to which the second pulse In a pair affects responding above this predicted level indicates the influence of transient system activity. 43

PAGE 51

44 The ratio of actual responding to predicted responding (based on I !near energy integration) to the second pulse II lustrates the relative dominance of transient or sustained system activity. If this ratio is greater than 1.0, transient activity is dominant, since the actual data exceed the predicted values. Ratios less than 1.0 ii lustrate sustained system dominance. For response amp I ltude, the ratios of actual to predicted (I lnear) response changes due to the presence of a second pulse In a pair are 7.28, 6.13, and 8.72, for Experiments 1, 2, and 3, respectively. For response probabll ity, these ratios are 5.33, 5.87, and 5.23, for Experiments 1, 2, and 3, respectively. Clearly, the effect of the second pulse In a pair Is much greater than would be expected If only I inear energy integration were occurring. The transient properties of the second pulse enhance responding, and transient system activity dominates sustained system activity when the two occur together. The response amp I itude data suggest that the result of temporal summation of transients Is more efficient than that of summation of sustained stimulus attributes, for both neonates and adults. However, the window of temporal summation Is narrower In adults (40 ms) than in infants (50 ms). Beyond this window, sustained summation becomes evident. Research using longer durations and intervals Is necessary to more accurately specify the transient summation window for neonates. The Infant and adult response probability data suggest that transient summation is maximal within 20 ms, and extends to at least 50 ms, with infant probabil lty being about 20 percent lower than adult probabll lty (see Figure 10), despite the fact that the stimulus was more Intense for infants than for adults. The temporal summation of

PAGE 52

C (l.) 0 L. (l.) Cl. >1-_J en 100 90 80 <( 70 en 0 0:: 0.. w 60 CJ) z 0 0.. CJ) 50 w 0:: 45 ADULT INFANT 3 20 35 50 65 80 100 INTERVAL (ms) Figure 10. Adult and infant response probability as a function of time for paired stimuli.

PAGE 53

46 transients fol lows the same time course in adults and neonates, for response probability. For adults, sustained summation contributed I ittle more than could be accounted for by transient summation, and this may have been due to a ceiling effect. In Infants, where maximal temporal summation of transients stil I yielded probabll itles below 70 percent, a clear additive effect of sustained summation was found above 20 ms. It may be the case that sustained summation would be more apparent in adults if stimuli which avoid the response probabil lty celling were used. When comparing infant and adult probabll ity for single stlmul i, the most striking finding is the much lower response probability for very brief stimuli in Infants (see Figure 11). Adults require much less stimulation than do infants to achieve equivalent response probabi I !ties. However, at longer durations single stimul I are quite effective in infants, supporting the belief that the sustained system is relatively mature In Infants (Graham, Strock, & Zeigler, 1981 ) Graham, Strock, and Zeigler ( 1981 ) suggest that transient immaturity may be due to both motor system and sensory system immaturity. The response of the orblcularis ocu Ii, the muscle responsible for the eyeb Ii nk, to a startle stimulus consists of two components, Rl and R2, which are neurologically dissimilar (Sanes, Foss, & Ison, 1982; Shahan! & Young, 1973). The Rl response Is a short latency (5-15 ms) response of motoneurons found in the palpebral segment of orbicular is ocul i. This Rl component occurs !psi laterally to electrical and tactile, but not auditory, stimul I. The R2 component has a longer latency (20-100 ms), and occurs in the orbital segment of orbicular is ocul I bilaterally. The neural circuit for R2 is more

PAGE 54

47 100 ADULTS ......... +C 90 INFANTS Q) u 'Q) a. ....... >80 I_J a)
PAGE 55

48 comp I teated than that for Rl, Involving brain stem nuclei at the level of the Inferior col I iculus. This R2 component is what Is measured as the startle eyebl ink by a periorbital electrode placement. For this reason, the R1 component may be Irrelevant to the present study. This means that the distinction between transient and sustained motoneurons may not be a relevant factor In defining the immaturity of the transient system, contrary to the suggestion of Graham, Strock, & Zeigler (1981). On the other hand, the suggestion by Graham, Strock, & Zeigler (1981) of a neurological dichotomy of transient and sustained sensory neurons is supported by the present research. The fact that pairs of stimuli can be as effective as analogous single stimuli, which contain much more energy, suggests that the Input to the startle center may be equivalent for the two types of stimul I. This means that temporal summation occurs either before or at the startle center, due to differential activation of transient and sustained sensory systems. In a previous report (Blumenthal & Berg, In press), It was suggested that startle amp I itude and probabil tty may be determined by partially independent underlying mechanisms, a startle trigger and a startle amp I ifier. This Is analogous to the distinction made in cognitive perceptual research between stimulus detection and stimulus identification (Posner, 1978). The extent to which the two stimuli in a pair of transients have common neural elements may have an effect on their summation. If the units have not fully recovered from the effects of the first pulse, then the contribution of the second pulse may be decreased. This is a type of forward masking, and the recovery of these units may be immature In neonates. This would explain the finding that response

PAGE 56

49 amp I itude increases as lnterstimulus interval increases above 20 ms for infants, but not for adults. This neural recovery may be more extended for the startle amp I ifler than for the trigger, since no evidence of increase in forward masking was found over the 20-50 ms range ot Intervals tested, for either Infant or adult response probability. The exact location and nature of the trigger and amplifier mechanisms is unknown, but they are most I ikely to be found either in the auditory pathway below the Interior col I lculus, or in the startle center Itself. Also, the two mechanisms may have different locations, since they appear to be partially independent in their responding. It either the trigger or the amp I lfier Is located in the auditory pathway, similar mechanisms may be found in other sensory pathways, and further studies Investigating startle trigger and amp I ifler mechanisms in visual and tactile modal !ties are suggested. The fact that analogous temporal summation abil ltles are found across modal lties suggests that this summation may be a very useful means of investigating startle trigger and amp I ifier mechanisms cross-modally. It would also be useful to conduct comparative studies using animals, to more accurately locate the mechanisms underlying these differential findings. Such studies might show that the trigger and amp I ifier mechanisms are due to specific startle center involvement, or to basic organizing principles found in al I sensory systems. In adults, sensitivity to transients may be greater tor the trigger than for the amp I lfler, since response probabil lty peaks at 20 ms for both single and paired stimuli, while the amp I itler continues to summate until 40 ms. This imp Iles that the amount of energy required to reach maximal activation is lower for the trigger than for the

PAGE 57

50 amp I if ier. In fact, this critical level may be below 20 ms, and further research at durations and Intervals of less than 20 ms Is needed to measure the transient sensitivity of the startle trigger. In infants, the immature transient sensitivity of the startle trigger is ii lustrated by the low response probabilities for brief duration stimuli (see Figure 11). The immaturity of the startle trigger is further ii lustrated by the fact that pairs of pulses are much less effective in Infants than in adults (see Figure 10). However, the infant and adult curves are parallel, so it may be the case that interstimulus interval is not the crucial parameter. The probability data suggest that both the single and paired stimulus functions show age differences which are due to Immature responding to very brief stimuli, regardless of the interstimulus interval. Making individual pulses in each pair longer, so that the critical energy could be reached at the shortest interval, Is necessary to separate the effects of increasing stimulus duration and lnterstimulus interval. Future research using pairs of longer stlmul i ls suggested, to separate these duration and interstimulus interval effects more fully. The present data support Graham and coworkers (1981) in their belief that the transient system is irrvnature in young infants. The infant response is immature In three ways. First, the window of temporal summation of transients, as ii lustrated by response amp I ltude, Is wider in infants than In adults. Second, the degree to which brief stimuli elicit the startle response, as ii lustrated by response probabil lty, is more pronounced in adults than in infants. Third, pairs of brief stimul I are less effective in el lcltlng transient system activation at al I Intervals In infants than in adults, al lowing

PAGE 58

51 transient and sustained summation to occur together in Infants, for response probabil lty.

PAGE 59

REFERENCES Anders, T., Emde, R., & Parmelee, A. A. (1971) A Manual of Standardized TermlnoloQy, TechnlQues. and Criteria for ScorlnQ of States of Sleep and Wakefulness In Newborn Infants. UCLA Brain Information Service, Los Angeles, CA. Apgar, V. (1953) Proposal for a new method of evaluation of the newborn Infant. Anesthesia and AnalQesla, .32., 260-267. Bartley, s. H. (1969) Principles of Perception, 2nd Edition. New York: Harper & Row. Berg, K. M. (1973) El icltatlon of acoustic startle in the human. (Doctoral dissertation, University of Wisconsin, 1973) Dissertation Abstracts International, 2A, 5217B-5218B. Berg, K. M., Berg, W. K., & Graham, F. K. (1971) Infant heart rate response as a function of stimulus and state. PsychophysloloQy, a, 3044. Blumenthal, T. D., & Berg, W. K. (1982) Developmental differences in startle response to stimulus onset transition. PsychophysloloQy, .19., 551-552 (abstract). Blumenthal, T. D., & Berg, W. K. (1984) Differential effects of transient and sustained components of acoustic stimul I on startle bl ink. Psychophysiology, .2..1, 569 (abstract). Blumenthal, T. D., & Berg, W. K. (In press) Stimulus rise time, intensity, and bandwidth effects on acoustic startle amp I itude and probability. Psychophysiology. Bohl In, G., Graham, F. K., SIiverstein, L. D., & Hackley, S. A. (1981) Cardiac orienting and startle bl Ink modification In novel and signal situations. Psychophysiology, J..a, 603-611. Brown, J. H. (1966) Magnitude estimation of angular velocity during passive rotation. Journal of Experimental Psychology, 12, 169-172. Brown, J. H. (1968) Cross-modal estimation of angular velocity. Perception and Psychophysics, 2, 115-117. Davis, M., & Gendelman, P. M. (1977) Plasticity of the acoustic startle response in the acutely decerebrate rat. Journal of Comparative and Physlologlcal Psychology, .21, 549-563. 52

PAGE 60

53 Davis, M., Gendelman, D.S., Tischler, M. D., & Gendelman, P. M. (1982) A primary acoustic startle circuit: Lesion and stimulation studies. Journal of Neuroscience, 2., 791-805. Davis, M., & Heninger, G. R. (1972) Comparison of response plasticity between the eyebl Ink and vertex potential in humans. Electroencephalography and Cl infcal Neurophysiology, .3..3., 283-293. Dykman, B. M., & Ison, R. (1979) Temporal integration of acoustic stimulation obtained In reflex Inhibition In rats and humans. Journal of Comparative and Physiological Psychology.~, 939-945. Fleshier, M. (1965) Adequate acoustic stimulus for startle reaction In the rat. Journal of Comparative and Physiological Psychology, .6.0., 200-207. Fox, J.E. (1979) Habituation and pre-stimulus inhibition of the auditory startle reflex in decerebrate rats. Physiology and Behavior, 23., 291297. Gal lager, D. W., Kehne, J. H., Wakeman, E. A., & Davis, M. (1983) Developmental changes In pharmacological responsivity of the acoustic startle reflex: Effects of picrotoxin. Psychopharmacology, l!l, 87-93. Gersunl, G. V. (1971) Temporal organization of the auditory function. In G. V. Gersuni (Ed.), Sensory Processes at the Neuronal and Behavioral Levels. New York: Academic Press. Geschelder, G. A. (1976) Evidence in support of the duplex theory of mechanoreception. Sensory Processes, 1, 68-76. Graham, F. K. (1975) The more or less start! Ing effects of weak prestimulatlon. Psychophysiology, 12, 238-248. Graham, F. K. (1979) Distinguishing among orienting, defense, and startle reflexes. In H. D. Kimmel, E. H. van 01st, and J. F. Orlebeke (Eds.), The Orienting Reflex in Humans.Hillsdale, NJ: Lawrence Erlbaum Associates. Graham, F. K., Anthony, B. J., & Zeigler, B. L. (1983) The orienting response and developmental processes. In D. Siddle (Ed.),~ Orienting Response. New York: J. Wiley and Sons. Graham, F. K., & Murray, G. M. (1977) Discordant effects of weak prestlmulation on magnitude and latency of the reflex bl Ink. Physiological Psychology, 2, 108-114. Graham, F. K., Strock, B. D., & Zeigler, B. L. (1981) Excitatory and inhibitory influences on reflex responsiveness. In W. A. Col I Ins (Ed.), Aspects of the Development of Competence.Hillsdale, NJ: Lawrence Erlbaum Associates. Green, D. M. (1973) Minimum Integration time. In A. R. Moller (Ed.),~ Mechanisms In Hearing. New York: Academic Press.

PAGE 61

54 Harbin, T. J., & Berg, W. K. (1983) The effects of age and prestlmulus duration upon reflex inhibition. Psychophysiology, .2Q, 603-610. Hickey, T. L. (1977) Postnatal development of the human lateral genlculate nucleus: Relationship to a critical period for the visual system. Science, 12.8., 836-838. Hlgashiyama, A., & Tashiro, T. (1983) Temporal and spatial Integration for electrocutaneous stimulation. Perception and Psychophysics, n, 437442. Hirsch, H. V. B., & Leventhal, A.G. (1978) Functional modification of the developing visual system. In M. Jacobson (Ed.), Handbook of Sensory Physiology, V.9: Development of Sensory Systems. New York: Springer Verlag. Hoffman, H. S., & Ison, J. R. (1980) Reflex modification in the domain of startle: 1. Some empirical f lndlngs and their imp I !cations for how the nervous stystem processes sensory input. Psychotogicat Review, .81, 175189. Kletsky, E. J., & Stengrevlcs, J.M. (1972) Estimation of the integration time-constant In auditory receptor units. Journal of the Acoustical Society of America, 22, 141 (abstract). Lacey, B. C., & Lacey, J. I. (1974) Studies of heart rate and other bodily processes in sensorimotor behavior. In P.A. Obrist, A.H. Black, J. Brener, and L. V. DiCara (Eds.), Cardiovascular Psychophysiology: Current Issues in Response Mechanlsms,Biofeedback and Methodology. Chicago, IL: Aldlne. Landis, C., & Hunt, W. A. (1939) The Startle Pattern. New York: Farrar and Rinehart. Marks, L. E. (1974) Sensory Processes. New York: Academic Press. Marsh, R., Hoffman, H. S., & Stitt, C. L. (1973) Temporal integration in the acoustic startle reflex of the rat. Journal of Comparative and PhysiotogJcat Psychology, .8.2, 507-511. Norman, J. L., Pettigrew, J. D., & Daniels, J. D. (1977) Early development of X-cel Is In kitten lateral geniculate nucleus. Science, 12.8., 202-204. Posner, M. I. (1978) Chronometric Explanations of Mind.Hillsdale, NJ: Lawrence Erlbaum Associates. Prokasy, W. F., & Ebel, H. C. (1967) Three components of the classically conditioned GSR In human subjects. Journal of Experimental Psychology, 12, 247-256. Raab, D. H. (1961) Forward and backward masking between acoustic cl lcks. Journal of the Acoustical Society of America, n, 137-139.

PAGE 62

55 Radionova, E. A. (1971) Two types of neurons in the cat cochlear nuclei and their role In audition. In G. V. Gersuni (Ed.), Sensory Processes at the Neuronal and Behavioral Levers. New York: Academic Press. Richards, A. M. (1976) Basic Experimentation io Psychoacoustics. Baltimore, MD: University Park Press. Sanes, J. N., Foss, J. A., & Ison, J. R. (1982) Conditions that affect the thresholds of the components of the eyebl Ink reflex In humans. Journal of Neurology, Neurosurgery, and Psychiatry .42, 543-549. Scharf, B. (1978) Loudness. In E. C. Carterette and M. P. Friedman (Eds.), Handbook of Perception, Y,4; Hearing. New York: Academic Press. Schwartz, S. H., & Loop, M. S. (1984) Effect of duration on detection by the chromatic and achromatic systems. Perception and Psychophysics, .3..6., 65-67. Shahan I, B. T., & Young, R.R. (1973) Bl ink reflexes In orbicular is ocul I. In J.E. Desmedt (Ed.), New Developments lo Electromyography and Cl ioical Neurophysiology, Yof. 3. Basel, Switzerland: S. Karger. Smal I, A. M., Jr., Brandt, J. F., & Cox, P. G. (1962) Loudness as a function of signal duration. Journal of the Acoustical Society of America, 24, 513-514. Stevens, J.C., & Hal I, J. W. (1966) Brightness and loudness as functions of stimulus duration. Perception and Psychophysics, 1, 319-327. Stevens, S.S. (1966) Duration, luminance, and the brightness exponent. Perception and Psychophysics, 1, 96-100. Stitt, C. L., Hoffman, H. S., & Marsh, R.R. (1976) Interaction versus Independence of startle-modification processes in the rat. Journal of Experimental Psychology: Animal Behavior Processes, 2, 260-265. Verril lo, R. T. (1968) A duplex mechanism of mechanoreception. In D.R. Kenshalo (Ed.), The Skin Senses. Springfield, IL: Charles C. Thomas Pub Yamada, F. (1983) Effects of intensity, duration, and rise and fal I times of the reflex evoking sound upon auditory evoked EMG of orblcularls ocul I and subjective estimation of startle. Japanese Journal of Physiological Psychology and Psychophysiology, 1, 11-18. Zeigler, B. L. (1978) Acoustic-startle modification by sustained and transient components of long lead stlmul i In human infants and adults. Unpubl I shed Master's thesis, University of Wisconsin Madison.

PAGE 63

56 Zeigler, B. L., Strock, B. D., and Graham, F. K. (1979) Facilitatory and inhibitory reflex preconditioning in early infancy. Paper presented at the Society for Research in Child Development Meetings, San Francisco, March, 1979. Zwislocki, J. J. (1960) Theory of temporal auditory summation. Journal of the Acoustlcal Society of America, .32, 1046-1060. Zwlslocki, J. J. (1969) Temporal summation of loudness: An analysis. Journal of the Acoustical Society of America, AQ, 431-441. Zwislocki, J. J. (1978) Masking: Experimental and theoretical aspects of simultaneous, forward, backward, and central masking. In E. C. Carterette and M. P. Friedman (Eds.), Handbook of Perception, Y,4; Hearing. New York: Academic Press. Zwlslockl, J. J., & Sokol ich, W. G. (1974) On loudness enhancement of a tone burst by a preceding tone burst. Perception and Psychophysics, 16., 87-90.

PAGE 64

BIOGRAPHICAL SKETCH Terry D. Blumenthal was born in Edmonton, Alberta, Canada, on July 17, 1954. He grew up In and around Roi ly View, Alberta, and eventually obtained a Bachelor of Science degree, with a specialization In psychology, at the University of Alberta in Edmonton. With this feat behind him, he decided to move from the top of a large pile to the bottom of a smaller pile, and he applied for graduate school. The University of Florida retained his services for five years, during which time he enjoyed I lfe to the fullest, with considerable help from his friends and tel low graduate students In a variety of departments. While at the University of Florida, he worked under the supervision of Dr. W. Keith Berg, who had a profound influence on how Mr. Blumenthal viewed the world. Five years after his arrival in Gainesvll le, Mr. Blumenthal moved on to a teaching position at Hamilton College in upstate New York, where he taught courses, wrote his doctoral dissertation, and had a very productive year on al I fronts. At this writing, matrimony looms on the horizon. Mr. Blumenthal has been consistently reminded of the considerable rewards which fol low hard work, personal appl !cation, and a wel I-told joke. 57

PAGE 65

I certify that I conforms to acceptable adequate, in scope and Doctor of Philosophy. have read this study and that in my opinion it standards of scholarly presentation and is fully quality, as a dissertation for the degree of &, --+-----'~-------------,,1w. Keith Berg, Chair Associate Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. W:rofessor of Speech I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my oprn10n it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Patricia H. Mille~ Associate Professor of Psychology

PAGE 66

I certify that I have read this study and that in my op1n1on it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Professor of Psychology This dissertation was submitted to the Graduate Faculty of the Department of Psychology in the College of Liberal Arts and Sciences and to the Graduate School, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August 1985 Dean, Graduate School

PAGE 67

lllll llllll l~~l1[ij1lirn1~~ ~m~rn11111111111 3 1262 08285 260 8