Development of Sound Localization 2. How do the neural mechanisms subserving sound localization develop?

Transcription

1 Development of Sound Localization 2 How do the neural mechanisms subserving sound localization develop? 1

2 Overview of the development of sound localization Gross localization responses are observed soon after the cochlea begins to function and in newborn humans. The precision of sound localization improves between birth and 5 years of age. Localization under complex listening conditions takes longer to develop. Experience appears necessary for the formation of auditory spatial maps. 2

3 Overview of this lecture Electrophysiological evidence of development of binaural hearing mechanisms in humans. Morphological and physiological evidence of development of binaural hearing mechanisms in nonhumans. Limitations imposed by immature peripheral coding. Development of spatial maps and role of experience. 3

4 ABR binaural interaction component McPherson et al. (1989) measured what is called the binaural interaction component of the ABR. This series of waveforms illustrate how the binaural interaction component is recorded. The top two waveforms are auditory brainstem responses recorded to clicks presented to the right ear (top) and the left ear (second from top). The third waveform from the top is the ABR recorded when both ears are stimulated simultaneously (binaural response). The monaural responses should represent neural responses along the auditory brainstem pathway, including those from some nuclei that are constructed to process interaural differences. But at least some neurons in those nuclei respond more when they are stimulated by both ears, and that neural response will only be reflected in the binaural ABR. So the difference between the sum of the two monaural responses (fourth waveform from top) and the binaural response represented this extra binaural response (bottom waveform). The difference between the summed monaural and the binaural responses in called the binaural interaction component. McPherson et al measured the BIC in adults and in newborn infants. The results for infants are shown on the right. Although he ABR waveform is clearly different in the newborn compared the adult, the BIC is quite similar, suggesting that binaural processing at the level of the brainstem is working. Given that newborns can tell whether a sound is coming from the left or right, this isn t very surprising. McPherson, D. L., Tures, C., & Starr, A. (1989). Binaural interaction of the auditory brain-stem potentials and middle latency auditory evoked potentials in infants and adults. Electroecepalogr. Clin. Neurophysiol., 74,

5 MLR binaural interaction component McPherson et al. also recorded binaural middle latency responses (MLR). The middle latency response is complex, in that it includes components that come from the thalamus as well as the reticular formation, which is involved in arousal and regulation of behavioral state. It can be difficult to record in infants because it is affected by state and sleep stage. Nonetheless, McPherson et al. were able to record the response in some newborns, to calculate a BIC in the same way that they calculated the ABR BIC. Their results are shown here, the adult BIC is at the bottom left, the infant on the bottom right. Infants show a BIC in the MLR, but it is different from that seen in adults- it looks like the polarity of the response is reversed in the infants. So this result suggests that the neural mechanism involved in binaural interactions at this level of the auditory system are immature at birth, but it is not clear exactly how they are immature. 5

6 Binaural responses detectable in most newborns Cone-Wesson et al. recorded the binaural interaction components for ABR Waves III and V and for the MLR in newborn infants. (ABR - 30 Hz HP refers to Wave V recorded with filter settings for MLR.) They used high and low level clicks as well as 500 and 4000 Hz tones to elicit responses. These researchers found that the majority of newborns produced BICs in all conditions. Cone-Wesson, B., Ma, E., Fowler, C.G. (1997). Effect of stimulus level and frequency on ABR and MLR binaural interaction in human neonates. Hear. Res., 106,

7 Newborn binaural responses suggest limitations on binaural processing Furst et al. (1990) examined the effect of ITD on ABR binaural interaction components. In adults they found that one of the waves of the BIC that they labeled β β Furst, M., Bresloff, I., Levine, R. A., Merlob, P. L., & Attias, J. J. (2004). Interaural time coincidence detectors are present at birth: evidence from binaural interaction. Hearing Research, 187(1-2), FURST M, EYAL S, KORCZYN AD PREDICTION OF BINAURAL CLICK LATERALIZATION BY BRAIN-STEM AUDITORY EVOKED- POTENTIALS HEARING RESEARCH 49 (1-3): NOV

8 Conclusion Binaural evoked potentials have not been well described in human infants 8

9 Morphological and physiological evidence of binaural development in nonhumans What limits binaural processing during development? The question, as before, is whether it is the acoustics, peripheral coding, or neural development that limits sound localization during development. I studies of nonhumans, as in the BIC ABR studies, the acoustics are irrelevant as the stimuli are presented under earphones and the acoustics are controlled. For a number of years, researchers who studied auditory neural development argued that the brainstem neural structures involved in sound localization were pretty much formed and ready to function as soon as the ear could give them some input. 9

10 Lateral superior olive: IID circuit Remember that the circuit involved in the earliest stage of interaural intensity difference processing involves bilateral projections from bushy cells in the anteroventral cochlear nucleus (AVCN) to the lateral superior olive (LSO). The contralateral AVCN projects to the medial nucleus of the trapezoid body (MNTB) and MNTB projects to the LSO, providing inhibit input to LSO neurons, while the ipsilateral AVCN projects directly to the LSO, providing an excitatory input. So LSO neurons respond when the sound is more intense at the ipsilateral ear. Brugge and his colleagues investigated the anatomy and physiology of this pathway in newborn kittens-- right at the time when the cochlea first starts to function, a few days before intense sounds can evoke a cochlear response. The picture they provided of the structure of the AVCN-MNTB-LSO circuit is shown at the bottom of the slide. The end bulbs of held on the AVCN neurons aren t as well developed, and the connections between neurons are not as extensive, but the circuit is there. Blatchley, B. J., & Brugge, J. F. (1990). Sensitivity to binaural intensity and phase difference cues in kitten inferior colliculus. J Neurophysiol, 64(2), Brugge, J. F. (1983). Development of the lower brainstem auditory nuclei. In R. Romand (Ed.), Developments of Auditory and Vestibular Systems (pp ). New York: Academic Press. 10

11 Medial superior olive: ITD circuit A similar situation holds for the circuit involved in calculating interaural time differences. Here, AVCN bushy cells project bilaterally to the neurons of the medial superior olive, and while inputs from either ear leads to a response in the MSO neurons, these neurons respond best when they receive simultaneous inputs from the two ears. The structures in newborns are shown at the bottom of the picture. The connections aren t as extensive as in the adult, but the circuit is there. 11

12 Responses of LSO neurons to IID This story of remarkably mature interaural processing continues with consideration of the responses of LSO and MSO neurons. This figure shows the response of single neurons to interaural intensity differences. The response is plotted along the y-axis as a function of the IID (on the x-axis). The response in an adult cat is shown in the x;s, a 13-day-old kitten is shown in the unfilled symbols. In both cases, the neuron responds best to an IID just over 0 db, and the response goes down as the IID deviates from that value. If anything, the kitten s response is more finely tuned than the adult's. 12

13 Responses of MSO neurons to ITD These are responses of MSO neurons to interaural phase differences (in π radians, so zero means no phase or time differences and 1.0 would be 1Xπ, so the sound in the two ears is completely out of phase (time difference equal to half the period), and at 2π, they re back in phase. The symbols are the same as in the last slide. These neurons seem to respond best to a a phase difference that slightly favors the ipsilateral ear, and again, the response of the kitten neuron is very similar to the response of the adult neuron. 13

14 Normalized spike rate? Notice that in both of these figures, the response of the neurons is plotted as normalized spike rate. What that means is that the spike rate is expressed as a percentage of the maximum spike rate of that neuron. Why is that important? 14

15 Immature neurons don t respond much It s important because the base maximum spike rate is very different in an immature animal than in a mature animal. The graph on the left illustrates the response of auditory nerve fibers in cats of different ages, as a function of sound intensity, from the Walsh and McGee paper. There is a progressive increase in the maximum response rate with age. The rate at which the spike rate increases as the intensity increases gets steeper with age as well. The graph on the right shows the slope of these spike rate-intensity function sin the auditory nerve and in the cochlear nucleus as a function of age. Over the early weeks of postnatal life, there is a progressive increase in hat slope in both the auditory nerve and the cochlear nucleus. 15

16 Immature LSO provides less information about IID Sanes and Rubel (1998) took a different look at the development of the LSO response. The graph here shows the actual spike rate of LSO neurons of kittens and adult cats. If you look at how much kitten and adult cat LSO neurons are responding as the IID changes, you see that the kitten neurons have a very limited range of response compared to the adult. The table puts some numbers on this. Dynamic range refers to the range of IIDs over which the average neuron responses. It is bigger in adult cats tan it is in or day old kittens. If you calculate how much the spike rate changes for each 1 db change in IID ( IID resolution in the table), the number is smaller in the kittens than in the adult cats. So the kitten brain ahs much less information telling it that the position of a sound source has changed. Sanes, D. H., & Walsh, E. J. (1998). The development of central auditory function. In E. W. Rubel, R. R. Fay & A. N. Popper (Eds.), Development of the auditory system (pp ). New York: Springer Verlag. Sanes, D. H., & Rubel, E. W. (1988). The functional ontogeny of inhibition and excitation in the gerbil auditory brain stem. J Neurosci, 8, Sanes, D. H., & Rubel, E. W. (1988). The ontogeny of inhibition and excitation in the gerbil lateral superior olive. J. Neurosci., 8(2),

17 Range of IIDs eliciting a response increases with age. Sanes and Rubel also noted that the range of IIDs to which neurons in the LSO responded also increased as the cats got older. The neurons in the kitten LSO were all tuned to a narrow range of negative (ipsilateral ear higher intensity) IIDs, while adult LSO neurons responded to IIDs across the entire range. This suggests again that IID processing is not mature at this time, and at least part of the problem is the dynamic range of the auditory neurons that provide the input to LSO. 17

18 Immature phase locking will lead to poor ITD processing We know that the MSO has to have well phase-locked inputs from the two ears to be able to calculate ITDs on a µsecond scale. Although we don t have good data on this, it is clear that if the the precision of phase locking is increasing over the early weeks of the cat s life, then the MSO is not going to be able to do a very good job of providing information about the precise ITD. 18

19 Conclusions re: interaural cue calculation in the immature auditory system The circuits used in calculating interaural differences are in place when the cochlea starts to function. The immature responses of neurons that provide input to the superior olive limit interaural cue calculation. The neurons of the superior olive may also be immature, independent of their inputs. 19

20 Forming a map of auditory space ITD µs IID 4-2 db Spectral shape Intensity -5-6 db degrees visual angle in in azimuth - 5 degrees visual angle in elevation.6.6 meters meters away away The idea of a map of auditory space is really the idea that every position in space has associated with it an array of sensory characteristics: auditory, visual, somatosensory, olfactory (not sure about taste ), vestibular and kinesthetic. In other words, if I hear a sound with certain interaural differences, a certain spectral shape, with a certain intensity, then I know that if I point my eyes to a position specified in terms of the visual field, I will see something there. Similarly if I reach my arm out at a particular angle, I will feel something there-- and so on. And somewhere in the brain there is a sort of look up table with entries for every position in space. When I talked about the effects of experience on the development of sound localization, I suggested that a lack of binaural experience might disrupt the formation of this map. I want to talk briefly about some studies describing how this process might occur. To understand this, you need to think about the problem. 20

21 The auditory system is laid out by frequency and calculates auditory space Auditory scene buzz hum click ring Neural computation of auditory space Intensity X Frequency X Time representation in the ear 22, -7,.6 20, -10,.6-10, -20,.6 20, -20,.4 Calculated spatial representation in the brain The cochlea is organized by frequency (tonotopic organization). Two receptors (hair cells) that are close together in the cochlea respond to frequencies that are close together. The output of the cochlea is organized by frequency and this organization is maintained throughout the auditory pathway. The auditory system has neural circuits that take the output of the cochlea and compute the locations of sound sources in space. 21

22 The visual system is laid out spatially View Spatial representation on retina 22, -7,.6 Retinotopic representation in the brain 20, -10,.6-10, -20,.6 20, -20,.4 The visual system is different. The sensory surface--the retina--is spatially organized. In other words, two receptors that are near to each other in the retina respond to light in places in space that are close together. This retinotopic, spatial, organization is maintained throughout the visual pathway. Some computation is necessary, of course: Because the retina is 2 dimensional, and space is 3 dimensional, the brain still has to figure out the distance dimension, but at least the 2 dimensions are evident in the retinal image. 22

23 Visual and auditory spatial representations are superimposed Spatial representation in auditory pathway Intensity X Frequency X Time representation in the ear 22, -7,.6 20, -10,.6 Multimodal spatial representation in the brain Scene buzz hum ring -10, -20,.6 click 20, -20,.4 Spatial representation on retina Some place in the brain, these two spatial representations are superimposed. The auditory system processes the sound, the visual system processes the light, and then somewhere in the brain, we find neurons that respond to places in space, whether the input from that place is sound or light. We say that representation is multimodal because it coordinates information from multiple modalities. One structure in which we find such a representation of space is in the superior colliculus. Physiological studies of the superior colliculus demonstrate that normal visual and auditory experience is necessary for the spatial map to develop normally. 23

24 Normal development of SC response in guinea pigs Azimuthal plane Neurons respond to sounds in these locations This slide shows how the spatial map in the superior colliculus (SC) develops in guinea pigs. A guinea pig s gestation period is days. Guinea pigs begin to hear prenatally. In terms of the cochlear response, a newborn guinea pig is similar to a 20-day-old kitten. So at this age, the response threshold at the level of the auditory nerve is pretty much like that of an adult. The figures at the right of this slide represent the auditory responses of neurons in the SC of 1-11 day-old guinea pigs (top) and adult guinea pigs (bottom). The big circles at the right of the figure represent the SC (anterior up, lateral to the right), and each dot is the location of one neuron that Withington-Wray et al. (1990) recorded from. Each of the smaller circles is a representation of the responses of one neuron and the letter next to the circle matches that neuron with the position labeled with the same letter in the schematic of the SC. The small circle is a representation of the azimuthal plane (the horizontal plane that passes through your ears). The figure within the circle outlines the positions in azimuth, in sound field, that that neuron would respond to. So, for example, the neuron labeled A was near the anterior end of SC and it responded to sounds that were located around 0º azimuth, The neuron labeled B was a little more posterior in SC than the one labeled A, and it responded to sounds that were located a little to the left of 0º azimuth. So notice two things: Each neuron has a limited section of auditory space to which it will respond and As you move from the anterior end of SC to the posterior end, the region to which the neuron responds shifts progressively counterclockwise. Notice that the neuron labeled G is the most posterior neuron and it responds to sounds at 180º azimuth. The responses of 1-11-day old guinea pigs shown at the top of the figure show neither of these characteristics. First, they all respond to sounds over rather large, unselective regions of auditory space. Second, as you move from the anterior end of SC to the posterior end, you don t see a systematic shift in the locations to which the neurons will respond. Withington-Wray et al. studied the development of the responses of these neurons and manipulated both auditory and visual experience to determine how experience influenced SC development.they either raised the animals in the dark, so that they had no visual experience or they raised the animals with constant omnidirectional noise. 24

25 Effects of visual and auditory experience on spatial maps Withington-Wray et al. studied the development of the responses of these neurons and manipulated both auditory and visual experience to determine how experience influenced SC development.they either raised the animals in the dark, so that they had no visual experience or they raised the animals with constant omnidirectional noise. 25

26 Effects of abnormal auditory experience on spatial maps These are the responses of neurons in the SC of animals who heard omnidirectional noise from birth on. Notice that they respond to a broad range of locations and that their best direction (if you can call it that) is not predicted by their location within SC. Withington-Wray, D. J., Binns, K. E., Dhanjal, S. S., Brickley, S. G., & Keating, M. J. (1990). The maturation of the superior collicular map of auditory space in the guinea pig is disrupted by developmental auditory deprivation. Eur J Neurosci, 2, Withington-Wray, D. J., Binns, K. E., & Keating, M. J. (1990). A four-day period of bimodality auditory and visual experience is sufficient to permit normal emergence of the map of auditory space in the guinea pig superior colliculus. Neurosci Lett, 116, Withington-Wray, D. J., Binns, K. E., & Keating, M. J. (1990). The developmental emergence of a map of auditory space in the superior colliculus of the guinea pig. Brain Res Dev Brain Res, 51, Withington-Wray, D. J., Binns, K. E., Dhanjal, S. S., Brickley, S. G., & Keating, M. J. (1990). The maturation of the superior collicular map of auditory space in the guinea pig is disrupted by developmental visual deprivation. Eur J Neurosci, 2,

27 Effects of dark rearing on spatial maps These figures show the effect of dark rearing on the auditory spatial responses of neurons in the SC. Each graph plots the location of an individual neuron in the SC as a function of its best direction response (the angle in azimuth at which that neuron responds the best). The results for normally reared day-old guinea pigs are in the rightmost figure. As you move through the SC the angle at which the neuron responds best shifts systematically from 0º azimuth to 180º azimuth. The results for 1-15-day-old normally reared animals are shown in the leftmost graph. These are like the results in the last slide. As you move through the SC, there is no systematic change in a neuron s best direction. The results for animals reared in the dark are shown in the middle panel. These animals are between 35 and 57 days of age, and like immature animals, they do not show a systematic spatiotopic organization. 27

28 Brief normal exposure is sufficient for normal spatial maps Withington-Wray et al. looked at groups of animals who were either dark reared or noise reared, except for brief periods--about 4 days long- - at different ages. And they found that only 4 days of normal visual and auditory inputs were enough to produce normal spatial maps, as long as the exposure was between 26 and 30 days. 28

29 Spectral as well as interaural cues are important These figures show the effects of removing an animals pinna-- preventing them from getting the spectral cues that allow us to distinguish front from back and locations in elevation. In these graphs position in the left hemifield is plotted as negative angles, while those in the right hemifield are shown in positive angles. The right SC responds to sounds (and sights) in the left hemifield. The normal organization of the SC responses to s sound are shown in the top left graph. As the position of the neuron in SC moves toward the posterior end of SC, the best direction shifts from 0º to 180º in azimuth. Animals who have had both pinnas removed early in life have SC units that respond to sounds all over the place-- both hemifields even. If you remove the pinnas of an adult animal, the neurons lose their ability to figure out the direction, but they all still respond to sounds in the left hemifield. Similarly, if the left pinna is removed, units in the SC on both sides of the brain tend to only respond to sounds in the left hemifield. Schnupp JWH, King AJ, Carlile S Altered spectral localization cues disrupt the development of the auditory space map in the superior colliculus of the ferret JOURNAL OF NEUROPHYSIOLOGY 79 (2): FEB

30 Abnormal experience can produce unusual neural responses. Normal experience Disparate experience Researchers have also shown that if inputs in the visual modality are shifted, using prisms, for example, then the auditory map shifts to match the visual map. So a neuron that would normally respond best to 0º azimuth will respond best to 15º azimuth, if prisms are used to shift the visual image 15º the the right. A recent study by Wallace and Stein suggests that during development, the auditory-visual map is very plastic. They dark reared kittens, except that every so often they would present simultaneous visual and auditory stimuli-- the trick was that the visual stimulus was in one location, while the auditory stimulus cam from a different location with no visual information associated with it. Some of the results of the study are shown in the graph on the right. The shaded areas represent locations to which neurons in the SC would respond (receptive fields). The left panel shows auditory (A, green) and visual (V, blue) receptive fields of neurons in a normally reared animal. The thing to notice is that the two receptive fields mostly overlap. The right panel shows the receptive fields of neurons in the animals who saw and heard things from different places simultaneously. They have receptive fields, but they have little overlap. So presumably, a cat reared in this way would run away from his mother instead of toward her if she meowed. Wallace MT, Stein BE Early experience determines how the senses will interact JOURNAL OF NEUROPHYSIOLOGY 97 (1): JAN

31 Where does experience have its effects? Spatial representation in auditory pathway Intensity X Frequency X Time representation in the ear 22, -7,.6 20, -10,.6 Multimodal spatial representation in the brain Scene buzz hum ring -10, -20,.6 click 20, -20,.4 Spatial representation on retina The measurements I have described have all been in the SC. That doesn t mean that experience doesn t affect the stages of processing prior to SC. It is known that the response of the ear isn t changed in animals exposed to omnidirectional noise during development, but ITD processing in the medial superior olive is affected. Seidl AH, Grothe B Development of sound localization mechanisms in the Mongolian gerbil is shaped by early acoustic experience JOURNAL OF NEUROPHYSIOLOGY 94 (2): AUG

32 Implications: Blind people and sound localization Blind people (and visually deprived guinea pigs) have same discrimination-type sound localization abilities as sighted people. Interestingly, they are able to localize sound sources by pointing as well as sighted people. Conclusion: Vision isn t the only sense that can define space. The studies of guinea pigs all show that if you expose adult animals to the same conditions that affected the spatial map in young animals, adult animals spatial maps are not affected. But does that mean that once the animal has been visually deprived, he will never be able to localize sounds? Various studies show that blind people can discriminate ITD and IID and spectral differences and distance based on auditory cues. They can tell if the position of a sound source changes. These results are not surprising-- we know that people with abnormal auditory experience can do these things, even though they can t point to the location of a sound source. But Zwiers et al (2001) also showed that blind people can point their nose or their arm to the position of a sound source as well as sighted people. What this suggests is that vision isn t the only sense or system that we use to define space-- there is somatosensory space and motor/kinesthetic space and coordination of the auditory system with those functioning systems allows for the development of auditory space. Pascual-Leone, A., Amedi, A., Fregni, F., Merabet, L. B., & On. (2005). The plastic human brain cortex. [Review]. Annual Review of Neuroscience, 28, Zwiers, M. P., Van Opstal, A. J., Cruysberg, J. R. M., & Ee. (2001). Two-dimensional sound-localization behavior of early-blind humans. [Article]. Experimental Brain Research, 140(2),

33 Conclusions Not surprisingly, binaural evoked responses can be evoked from newborn infants, although the morphology of some responses change with age Whether binaural interaction or improvements in monaural coding is responsible for changes in response is not clear. Normal multimodal experience is required for the formation of auditory maps of space. 33