Interview with David Stapells Ph.D., Professor and Director, School of Audiology and Speech Sciences, University of British Columbia, Vancouver, BC Canada
Share:
Topic: Auditory Steady State Response (ASSR)
Beck: Good morning David. Thanks for your time today.
Stapells: Hi Doug. Thanks for the invitation.
Beck: David, if you don't mind, before we get to ASSR, please tell me a little about your professional background. I understand you're not an audiologist?
Stapells: No, I'm not, though I've been working with audiologists for more than twenty years. My bachelor's is in Psychology, and my Ph.D. is in Sensory Psychology, a branch of Experimental Psychology. Basically, that means I focused on hearing and related evoked potentials through my doctoral work. I earned my Ph.D. in 1984 and then went to UCSD and did a two-year post-doc with Dr Robert Galambos to continue my research into ASSRs, mostly at the Children's Hospital in San Diego. After that I moved to Albert Einstein College of Medicine in the Bronx, and spent some 10 years there running the Auditory Evoked Potential Laboratories. We did both clinical ABR work and AEP research -- it was here that I obtained my clinical experience and worked out solutions to clinical problems through my research. In 1996, I returned home to Vancouver, where I've been since at the University of British Columbia as a professor in the School of Audiology and Speech Sciences.
Beck: David, please tell me the best possible working definition of ASSR?
Stapells: Good question. I think one can think about ASSR as a repetitive evoked potential which is best considered in terms of its constituent frequency components rather than in terms of its waveform. If stimulus rates are high enough, the resulting response often resembles a sinusoidal waveform whose fundamental frequency is the same as the stimulation rate.
Beck: And just to be clear, the steady-state response does not have to resemble a sinusoidal response?
Stapells: You're correct. The ASSR has been incorrectly characterized as always being a sinusoidal response, and that is just not correct - it could be a more complex waveform and not have a sinusoidal appearing response. Typically, however, when the repetition rate is fast enough, the major frequency components are actually at the rate of stimulation.
Beck: What can you tell me about commercially available systems and automatic detection of the ASSR?
Stapells: Let me start by saying that the two or three ASSR detection algorithms I am familiar with are quite good and have been proven in well-designed studies to do what they purport to doing. Some of the clinical ASSR systems use these proven response detection algorithms and stimulus paradigms. Other systems, however, employ detection or stimulus paradigms that do not have a published literature. These need to be researched. The adequacy of response detection algorithms is all the more important because they remove the human from the system -- mistakes can be made which are difficult to determine after the fact. Of course, the opposite is true too, and there are always opportunities for error, with and without human oversight. Currently, perhaps the best thing is to have the machine detect the ASSR with confirmation using well-tested methods requiring human input, namely tone-evoked ABR and behavioral audiometry.
Beck: David, I know that in the lectures I've attended on ASSR, stimulus artifact can be an issue, especially with the higher intensity stimuli. Can you address stimulus artifact for me?
Stapells: Yes. Stimulus artifact needs to be watched for. With the ABR, the stimulus artifact is separate from the response, so the human observer can separate these. With computer-detected ASSRs, this is not the case. One issue is that if the ASSR system does not use the right analog-to-digital conversion set-up, you can get "aliasing," which could also be an issue with digital hearing aids, especially if you don't sample at a high enough rate. My Ph.D. student, Susan Small, has shown that there are spurious ASSRs due to aliased stimulus artifact that can occur when using high-intensity air- or bone-conducted stimuli, particularly with severe-to-profound hearing loss patients. Additionally, another potential source for spurious ASSRs might be vestibular or vestibular myogenic responses. Generally, we don't see the vestibular response in patients with normal hearing, or even a mild or moderate response, because their auditory responses are so large. But, the vestibular response could be present in an ASSR from patients with severe-to-profound hearing loss.
Beck: Interesting thought. So how can you separate out vestibular from auditory responses?
Stapells: Well, we don't know just yet. We do know that these responses occur essentially with low-frequency stimuli, and we have not seen it at 2000 and 4000 Hz. But, it is difficult to know if the ASSR has a vestibular or auditory component. We have methods to be quite certain that the response we're seeing is, indeed, physiologic, but whether it is from the vestibular or the auditory system, when using very loud presentations, is still unknown, and this is a major concern for all of us.
Beck: I'm curious to know why when you first observed this possible vestibular response, you thought it might be vestibular in origin?
Stapells: Well, we knew from the response morphology characteristics that it was physiologic, as opposed to electrical. As you know Doug, with ABR you can alternate the stimuli to get rid of the artifact sometimes, and we tried that, but it didn't get rid of every response. Alternating the stimulus did get rid of the responses due to aliasing. We used some other techniques too, but it stayed...so we believe it is physiologic in origin. We actually do not know for sure that it is vestibular, or vestibular myogenic or something else, we're just not sure.
Beck: Another issue with ASSR is the clinician actually just looks at the overall pattern of response, rather than the actual factors such as latency or amplitude...is that right?
Stapells: Yes, more or less.... In general, audiologists look to see if there is a response, and if so, it is assumed that the ear has been stimulated and created a response. Maybe if we also looked at the raw amplitude and latency (actually, phase) data, we would be able to get more out of the responses, that is, maybe the raw numbers would prove to be more useful, or perhaps less ambiguous, especially for possibly spurious responses.
Beck: Can you address the issue of using the ASSR as a screening tool for newborns?
Stapells: Great question. Many labs around the world, including our lab, are currently investigating using the ASSR for rapid newborn screenings.
I know there are at least three companies investigating automatic screening ASSR tests...but one problem is defining when these products
are actually ready for massive release. In other words, we need to know that positive results are really positive, and negative results are really negative. Large-scale research trials would be needed. ABR and OAE tools are very reliable and have excellent history demonstrating their efficacy as screening tools, so they have set a rather high target, and that's ultimately a very good thing. However, in a recent study, it was determined that ASSR is not a very good test for newborns, so at this point, using ASSR for a screening tool is interesting, but not yet ready for massive implementation in newborn screenings.
Beck: What would be the advantage in using ASSR, rather than ABR and OAE with regard to screenings?
Stapells: Well, there are a number of potential advantages. The ASSR could actually test both ears simultaneously, whereas ABR and OAE systems really just test one ear at a time, or they switch rapidly, but still, it's one ear at a time. Another issue is that ASSR might give us relatively quick frequency-specific information, and that would be very useful for counseling, fitting hearing aids, for observing changes over time, and certainly other functions would be defined, too. So there are a few advantages to ASSR screenings, if we can make them work as I think we will, but it'll take time to get there!
Beck: Can ASSR be performed via bone conduction?
Stapells: Yes. Unfortunately, it's still uncommon for people to do bone- conduction tests in electrophysiology. It's similar to ABR in that it is certainly possible, but few people do it, and those who do it well certainly need to address calibration issues, especially calibration of the bone conduction stimulator. We have recently collected infant data on bone conduction ASSR, and we will present these this summer, but there's nothing yet published on ASSR via bone conduction in infants with hearing loss, so it's ripe for analysis and research.
Beck: One of the issues you raised at the Phonak Pediatric Conference had to do with separating ASSR "marketing" from "science." Can you address that here?
Stapells: Sure. There have been claims I've heard, which I question. So without pointing a finger at anyone in particular, I'd say that when you hear claims such as "New ways of analyzing ASSR can do this, that or the other" without presenting evidence to back that -- that's a problem. Another claim is that the ASSR is "more frequency-specific than ABR." That's a problem too, because the only two studies that addressed this issue showed that the results were indeed good, but not different than a tone-based ABR. Lastly, I've also heard a claim that so-and-so's ASSR is better at assessing profound losses than is ABR, but that too, has never actually been properly studied. I have problems with claims which do not have appropriate scientific studies behind them. When you compare ABR and ASSR you have to compare apples to apples and oranges to oranges, and that hasn't been well done. You could indeed compare ABR clicks to ASSR clicks and then make a statement, or you could compare ASSR tones to ABR tones, but that too, has not been done appropriately. So, that's the issue, you need some equivalence in the stimuli to make claims of superiority.
Beck: Can we say that given equivalent protocols and stimulus, that a well-done tone-based ABR in excellent hands, and a well done tone-based ASSR in excellent hands, would determine essentially the same hearing thresholds?
Stapells: Yes, I would expect that they would have reasonably similar results for adults with sensorineural hearing loss.
Beck: David, this is fascinating discussion and I am very appreciative of our time today.
Stapells: Thanks Doug, it's been a pleasure speaking with you, too.
******************************************************
SPECIAL THANKS to PHONAK for facilitating this interview at their 2004 Pediatric Conference in Chicago.
Dr. Stapell's Website: www.audiospeech.ubc.ca/haplab/clinic.htm
Recommended Readings:
Herdman, A.T. & Stapells, D.R.(2003). Auditory steady-state response thresholds of adults with sensorineural hearing impairments. International Journal of Audiology:42, 237-248.
Picton, T. W., John, M. S., Dimitrijevic, A., & Purcell, D. (2003). Human auditory steady-state responses. International Journal of Audiology: 42(4), 177-219.
Small, S.A. & Stapells, D.R. (2004). Artifactual responses when recording auditory steady-state responses. Ear and Hearing: 25(6), 611-623.
Small, S.A. & Stapells, D.R. (2005). Multiple auditory steady-state response thresholds to bone-conduction stimuli in adults with normal hearing. Journal of the American Academy of Audiology: 16: 172-183.
Stapells: Hi Doug. Thanks for the invitation.
Beck: David, if you don't mind, before we get to ASSR, please tell me a little about your professional background. I understand you're not an audiologist?
Stapells: No, I'm not, though I've been working with audiologists for more than twenty years. My bachelor's is in Psychology, and my Ph.D. is in Sensory Psychology, a branch of Experimental Psychology. Basically, that means I focused on hearing and related evoked potentials through my doctoral work. I earned my Ph.D. in 1984 and then went to UCSD and did a two-year post-doc with Dr Robert Galambos to continue my research into ASSRs, mostly at the Children's Hospital in San Diego. After that I moved to Albert Einstein College of Medicine in the Bronx, and spent some 10 years there running the Auditory Evoked Potential Laboratories. We did both clinical ABR work and AEP research -- it was here that I obtained my clinical experience and worked out solutions to clinical problems through my research. In 1996, I returned home to Vancouver, where I've been since at the University of British Columbia as a professor in the School of Audiology and Speech Sciences.
Beck: David, please tell me the best possible working definition of ASSR?
Stapells: Good question. I think one can think about ASSR as a repetitive evoked potential which is best considered in terms of its constituent frequency components rather than in terms of its waveform. If stimulus rates are high enough, the resulting response often resembles a sinusoidal waveform whose fundamental frequency is the same as the stimulation rate.
Beck: And just to be clear, the steady-state response does not have to resemble a sinusoidal response?
Stapells: You're correct. The ASSR has been incorrectly characterized as always being a sinusoidal response, and that is just not correct - it could be a more complex waveform and not have a sinusoidal appearing response. Typically, however, when the repetition rate is fast enough, the major frequency components are actually at the rate of stimulation.
Beck: What can you tell me about commercially available systems and automatic detection of the ASSR?
Stapells: Let me start by saying that the two or three ASSR detection algorithms I am familiar with are quite good and have been proven in well-designed studies to do what they purport to doing. Some of the clinical ASSR systems use these proven response detection algorithms and stimulus paradigms. Other systems, however, employ detection or stimulus paradigms that do not have a published literature. These need to be researched. The adequacy of response detection algorithms is all the more important because they remove the human from the system -- mistakes can be made which are difficult to determine after the fact. Of course, the opposite is true too, and there are always opportunities for error, with and without human oversight. Currently, perhaps the best thing is to have the machine detect the ASSR with confirmation using well-tested methods requiring human input, namely tone-evoked ABR and behavioral audiometry.
Beck: David, I know that in the lectures I've attended on ASSR, stimulus artifact can be an issue, especially with the higher intensity stimuli. Can you address stimulus artifact for me?
Stapells: Yes. Stimulus artifact needs to be watched for. With the ABR, the stimulus artifact is separate from the response, so the human observer can separate these. With computer-detected ASSRs, this is not the case. One issue is that if the ASSR system does not use the right analog-to-digital conversion set-up, you can get "aliasing," which could also be an issue with digital hearing aids, especially if you don't sample at a high enough rate. My Ph.D. student, Susan Small, has shown that there are spurious ASSRs due to aliased stimulus artifact that can occur when using high-intensity air- or bone-conducted stimuli, particularly with severe-to-profound hearing loss patients. Additionally, another potential source for spurious ASSRs might be vestibular or vestibular myogenic responses. Generally, we don't see the vestibular response in patients with normal hearing, or even a mild or moderate response, because their auditory responses are so large. But, the vestibular response could be present in an ASSR from patients with severe-to-profound hearing loss.
Beck: Interesting thought. So how can you separate out vestibular from auditory responses?
Stapells: Well, we don't know just yet. We do know that these responses occur essentially with low-frequency stimuli, and we have not seen it at 2000 and 4000 Hz. But, it is difficult to know if the ASSR has a vestibular or auditory component. We have methods to be quite certain that the response we're seeing is, indeed, physiologic, but whether it is from the vestibular or the auditory system, when using very loud presentations, is still unknown, and this is a major concern for all of us.
Beck: I'm curious to know why when you first observed this possible vestibular response, you thought it might be vestibular in origin?
Stapells: Well, we knew from the response morphology characteristics that it was physiologic, as opposed to electrical. As you know Doug, with ABR you can alternate the stimuli to get rid of the artifact sometimes, and we tried that, but it didn't get rid of every response. Alternating the stimulus did get rid of the responses due to aliasing. We used some other techniques too, but it stayed...so we believe it is physiologic in origin. We actually do not know for sure that it is vestibular, or vestibular myogenic or something else, we're just not sure.
Beck: Another issue with ASSR is the clinician actually just looks at the overall pattern of response, rather than the actual factors such as latency or amplitude...is that right?
Stapells: Yes, more or less.... In general, audiologists look to see if there is a response, and if so, it is assumed that the ear has been stimulated and created a response. Maybe if we also looked at the raw amplitude and latency (actually, phase) data, we would be able to get more out of the responses, that is, maybe the raw numbers would prove to be more useful, or perhaps less ambiguous, especially for possibly spurious responses.
Beck: Can you address the issue of using the ASSR as a screening tool for newborns?
Stapells: Great question. Many labs around the world, including our lab, are currently investigating using the ASSR for rapid newborn screenings.
I know there are at least three companies investigating automatic screening ASSR tests...but one problem is defining when these products
are actually ready for massive release. In other words, we need to know that positive results are really positive, and negative results are really negative. Large-scale research trials would be needed. ABR and OAE tools are very reliable and have excellent history demonstrating their efficacy as screening tools, so they have set a rather high target, and that's ultimately a very good thing. However, in a recent study, it was determined that ASSR is not a very good test for newborns, so at this point, using ASSR for a screening tool is interesting, but not yet ready for massive implementation in newborn screenings.
Beck: What would be the advantage in using ASSR, rather than ABR and OAE with regard to screenings?
Stapells: Well, there are a number of potential advantages. The ASSR could actually test both ears simultaneously, whereas ABR and OAE systems really just test one ear at a time, or they switch rapidly, but still, it's one ear at a time. Another issue is that ASSR might give us relatively quick frequency-specific information, and that would be very useful for counseling, fitting hearing aids, for observing changes over time, and certainly other functions would be defined, too. So there are a few advantages to ASSR screenings, if we can make them work as I think we will, but it'll take time to get there!
Beck: Can ASSR be performed via bone conduction?
Stapells: Yes. Unfortunately, it's still uncommon for people to do bone- conduction tests in electrophysiology. It's similar to ABR in that it is certainly possible, but few people do it, and those who do it well certainly need to address calibration issues, especially calibration of the bone conduction stimulator. We have recently collected infant data on bone conduction ASSR, and we will present these this summer, but there's nothing yet published on ASSR via bone conduction in infants with hearing loss, so it's ripe for analysis and research.
Beck: One of the issues you raised at the Phonak Pediatric Conference had to do with separating ASSR "marketing" from "science." Can you address that here?
Stapells: Sure. There have been claims I've heard, which I question. So without pointing a finger at anyone in particular, I'd say that when you hear claims such as "New ways of analyzing ASSR can do this, that or the other" without presenting evidence to back that -- that's a problem. Another claim is that the ASSR is "more frequency-specific than ABR." That's a problem too, because the only two studies that addressed this issue showed that the results were indeed good, but not different than a tone-based ABR. Lastly, I've also heard a claim that so-and-so's ASSR is better at assessing profound losses than is ABR, but that too, has never actually been properly studied. I have problems with claims which do not have appropriate scientific studies behind them. When you compare ABR and ASSR you have to compare apples to apples and oranges to oranges, and that hasn't been well done. You could indeed compare ABR clicks to ASSR clicks and then make a statement, or you could compare ASSR tones to ABR tones, but that too, has not been done appropriately. So, that's the issue, you need some equivalence in the stimuli to make claims of superiority.
Beck: Can we say that given equivalent protocols and stimulus, that a well-done tone-based ABR in excellent hands, and a well done tone-based ASSR in excellent hands, would determine essentially the same hearing thresholds?
Stapells: Yes, I would expect that they would have reasonably similar results for adults with sensorineural hearing loss.
Beck: David, this is fascinating discussion and I am very appreciative of our time today.
Stapells: Thanks Doug, it's been a pleasure speaking with you, too.
******************************************************
SPECIAL THANKS to PHONAK for facilitating this interview at their 2004 Pediatric Conference in Chicago.
Dr. Stapell's Website: www.audiospeech.ubc.ca/haplab/clinic.htm
Recommended Readings:
Herdman, A.T. & Stapells, D.R.(2003). Auditory steady-state response thresholds of adults with sensorineural hearing impairments. International Journal of Audiology:42, 237-248.
Picton, T. W., John, M. S., Dimitrijevic, A., & Purcell, D. (2003). Human auditory steady-state responses. International Journal of Audiology: 42(4), 177-219.
Small, S.A. & Stapells, D.R. (2004). Artifactual responses when recording auditory steady-state responses. Ear and Hearing: 25(6), 611-623.
Small, S.A. & Stapells, D.R. (2005). Multiple auditory steady-state response thresholds to bone-conduction stimuli in adults with normal hearing. Journal of the American Academy of Audiology: 16: 172-183.
