NAME: Gloria Nyarko PROG.: EMCL+
ID NUMBER: 799847
A SHORT REPORT ON A 7-WEEK INTERNSHIP UNDERTAKEN AT THE SPEECH LAB The Speech Lab – University of Potsdam
The Speech lab is an ultramodern laboratory situated at the University of Potsdam’s Golm campus. The lab has equipment used in research of speech processing and production.
Equipment include a sound-proof booth which is ideal for making all sorts of recordings. Apart from the various projects undertaken by students and staff in Phonetics and Phonology, the lab is used by other departments. There has been for example a collaboration in the past with the baby lab for research in speech of children. Equipment found in the lab are also requested from time to time for recording of participants outside the lab by other departments. In my time spent in the speech lab, there was a request for a high quality audio recorder for recording of an endangered language. The department and the lab proves to be an invaluable asset to the Institution.
The presence of computer based tools for assessment of speech and language sparked my interest in working with the lab as I have a strong interest in technical tools and computers. I also found the research topic interesting because I was informed about the use of Praat (Boersma, 2001) for the later analysis of the acoustic data that will be collected. Praat was introduced to me at the University of Eastern Finland where intensive courses and exercises were done on it. It was therefore a great opportunity to use what I have learnt in a very practical situation. I also picked up a few tips from my daily supervisor on scripting as well as shortcuts that would make analysis a bit easier than usual. Apart from Praat, I had the opportunity to be introduced to Matlab and Audacity which were new to me.
Background to the study
The motor theory of speech perception (D’Ausilio et al., 2009b; Fadiga & Craighero, 2006;
Pulvermuller, Huss, Kherif, & Moscoso del Prado Martin, 2006) posits that recognition of speech sounds relies on mapping acoustic speech input onto motor speech representations in the listener (Liberman, 1957; Liberman, Cooper, Shankweiler, & Studdert-Kennedy, 1967; Liberman
& Mattingly, 1985). This implies that what a speaker hears influences what (s)he says. Consistent with this claim, functional imaging and transcranial magnetic stimulation (TMS) work has found that the motor speech system is indeed activated during speech listening (Watkins, Strafella, &
Paus, 2003; Wilson, Saygin, Sereno, & Iacoboni, 2004) and TMS interference of motor cortex has been reported to modulate speech perception under some conditions (D’Ausilio et al., 2009b;
Meister, Wilson, Deblieck, Wu, & Iacoboni, 2007; Watkins & Paus, 2004).
Prior to Hickok, Costanzo, Capasso & Miceli (2011), several experiments had also been conducted using language impaired populations to determine whether the motor system was crucial for language perception, i.e. if important for language perception, then damage to the areas of the brain associated with language production should result in language perception deficits. While some researchers concluded that an impairment in the production network caused certain perception difficulties (example Basso, Casati, & Vignolo, 1977; Baker, Blumsteim, & Goodglass, 1981), these claims were debated as their methods of assessment did not include benefits of modern day neuroimaging support to rule out multiple lesion sites and many others. Others that used well defined methods and concluded that the motor system was not crucial for language perception used small sample sizes that could not be said to be representative of the actual population (example Rogalsky, Love, Driscoll, Anderson & Hickok, 2011). Hickok et al. (2011), revisited the role of the motor speech system in speech perception in a larger series of patients with radiologically confirmed damage to the motor speech system defined minimally as damage involving Broca’s area. It was concluded from the study that the motor system was not critical for speech perception, however, performance in syllable discrimination was significantly worse than performance of control participants. This implied that damage involving Broca’s region may result in mild decline in performance on syllable discrimination.
In psycholinguistic experiments which have investigated the subject matter of the motor system’s impact on the perception of language, methods such as ERP and Eye Tracking have been used. A more recent addition to these is the use of auditory distractors in speech production tasks to determine whether perception indeed has an effect on production. In such experiments, the interest is in whether production time will be longer or shorter based on what is perceived and at what point in time.
Roon & Gafos (2014), in a replication study of theGalantucci, Fowler, & Goldstein (2009) study, investigated perceptuomotor effects of response-distractor compatibility in speech using a cue- distractor paradigm. This paradigm involves the presentation of a response cue, either visually or auditorilly, then shortly after, a presentation of a distractor, which is usually presented before the responses of the participants are planned and executed. The distance between the presentation of the response cue and the perceived distractor is known as the Stimulus Onset Asynchrony (SOA), which must be manipulated to establish that reaction time modulations unambiguously reflect an influence of perception on production.
Whereas Galantucci et al. aimed at matching on the phonemic level (i.e. whether syllables heard had same or different phonemes that differed by one phoneme in the same position), Roon &
Gafos (2014) in the replication study went beyond this to include in their study place of articulation and voicing congruency. SOA manipulation was employed in the investigation of perceptuomotor effects of voicing and articulator and was found that voicing and place of articulation do indeed affect target responses. If they shared the same qualities, reaction times were faster but slower if they did not.
The Current Study
The current study employed the use of the cue-distractor paradigm. The interest in this study is however not in consonants but in vowels. Evidence has been found that consonants and vowels are processed differently in some language impaired population (eg. Caramazza, Chialant &
Miceli, 2000; Boatman, Hall, Goldstein, Lesser & Gordon, 1997). Also, fMRI studies conducted
Based on these claims, a set of participants were recruited to participate in the ongoing study to determine whether features shared by vowels would or would not affect what is produced when heard. It is proposed that distractor effects of pairs with no phonological features in common will be greater than those with one or more features in common.
For this study, the cues were presented visually and then an auditory distractor was presented shortly after. To find the best times that a distractor would have an effect on participant responses, a pilot study was conducted and two SOAs, 150ms and 200ms, were proposed as being the best intervals for cue and distractor presentations. Both were used in the experiment.
Participants selected for the study were self-reported native German speakers. To begin, they sat in front of a computer and headsets were given to them to wear. They were instructed to say /ee/ when they saw ** and /uu/ when they saw ##. The distractors heard were not limited to just these two but also included /o/, /i/ and a tone as well. Participants were instructed to respond quickly, as soon as they saw the symbols on the screen, but not so quickly that they made too many errors. Participants were told to ignore what they heard during the experiment because it was irrelevant for the experimental task.
During the experiment, there were two breaks which allowed participants to take a rest from the experiment before they continued. The time for the experiment was approximately 30 minutes.
Psychtoolbox 3 implemented in Matlab was used in the presentation of the visual and auditory stimuli and responses of the participants, picked up by the microphones beside them, were recorded using Audacity.
Duties in the lab
Before commencement of the internship, data collection was already underway in the lab. The project was explained to me and the procedures for receiving participants and running experiments were also explained in detail. Just as was expected, a language background questionnaire as well as an informed consent form were presented to participants who had to fill and sign respective documents. After the first day, I received the participants and run the experiment unassisted for two weeks. There were daily reports to the supervisor about progress
Simultaneously, analysis of data was started using Praat (Boersma, 2001). The first step to this was to mark boundaries from point of cue presentation to the beginning of the response of the participant for all trials. These were rough boundaries. After this was done for all participants, I had to go over them again, this time using a praat script. Now zoomed in, I refined the boundaries at the cue by moving it to about 50% of the marker. As explained by the supervisor, this was so that it would be consistent throughout for all trials and for all participants. The next step was to click on the script and then a combination of the control key and R on the keyboard. This when done moves the selection to the next interval and then these steps were repeated.
Completed cue-refined data were forwarded to the supervisor who cross-checked and then started on the response refinement. He explained that he had to go through that because he had to determine whether the boundary should be placed before the actual production or just when participants were preparing to produce the vowels.
After all of these were completed, annotation of cues, distractor and response were done by the supervisor. However, the annotation in the cue tier were repeated in the response section as those are the responses expected from the participants. It was then my duty to listen to each response in each trial also using a praat script, moving from boundary to boundary and change where errors were made by the participants or where there was a yawn or no response or an inaudible response. These were excluded from the data analyses.
As at the time of completion of my 7th week on the project, it was established that more participants had to be recruited for more data as those collected contained too many errors. The preliminary findings were however consistent with the proposed hypothesis that distractor effects of pairs with no phonological features in common will be greater than those with one or more features in common. There were longer reaction times when participants heard vowels that shared no features with the target sounds than those that shared one two or three features.
Summary of my experiences in the speech lab
Even though this project is not in the direction of my final master thesis, I feel priviledged to have had the opportunity to undertake my internship at the speech lab and under the Head of Department for the Linguistics department. The project was well coordinated and the systematic
the experiment went, informed me adequately about how participants felt during the experiment. Some felt it was just okay, others thought it went too fast for them and needed more time to process one presentation before the other started. A participant advised me to partake in the experiment to know exactly how it felt because she felt I must do so to understand how she felt. A participant confessed that with the sounds played in her ear, she knew she was being tricked to make errors and she made a conscious effort to avoid them but somehow kept making them. These were comments I found very interesting and even more rewarding were the participants who wanted to know what the experiment was about. It was a very good opportunity to explain the different fields of Linguistics and the experiment they had participated in.
Acoustic data analysis I must say is tough. But it is a rare opportunity to get to do an actual acoustic analysis with a skill learnt that is not just class or homework.
REFERENCES
Baker, E., Blumsteim, S. E., & Goodglass, H. (1981). Interaction between phonological and semantic factors in auditory comprehension. Neuropsychologia, 19, 1–15.
Basso, A., Casati, G., & Vignolo, L. A. (1977). Phonemic identification defects in aphasia.
Cortex, 13, 84–95.
Boatman, D., Hall, C., Goldstein, M., Lesser, R., & Gordon, B. (1997). Neuroperceptual differences in consonant and vowel discrimination: As revealed by direct cortical electrical interference. Cortex; a Journal Devoted to the Study of the Nervous System and Behavior,33(1), 83-98.
Boersma, Paul (2001). Praat, a system for doing phonetics by computer. Glot International5:9/10, 341-345.
Caramazza, A., Chialant, D., Capasso, R., & Miceli, G. (2000). Separable processing of consonants and vowels. Nature, 403(6768), 428-30.
D’Ausilio, A., Pulvermuller, F., Salmas, P., Bufalari, I., Begliomini, C., & Fadiga, L. (2009b).
The motor somatotopy of speech perception. Current Biology, 19(5), 381–385.
Fadiga, L., & Craighero, L. (2006). Hand actions and speech representation in Broca’s area.
Cortex, 42(4), 486–490.
Galantucci, B., Fowler, C. A., & Goldstein, L. (2009). Perceptuomotor compatibility effects in speech. Attention, Perception, &Psychophysics, 71(5), 1138–1149.
Hickok G., Costanzo M., Capasso R. , Miceli R. (2011). The role of Broca’s area in speech perception: Evidence from aphasia revisited. Brain and Language, 119 (3), 214-220.
Liberman, A. M. (1957). Some results of research on speech perception. Journal of the Acoustical Society of America, 29(1), 117–123.
Liberman, A. M., Cooper, F. S., Shankweiler, D. P., & Studdert-Kennedy, M. (1967).
Perception of the speech code. Psychological Review, 74(6), 431–461.
Liberman, A. M., & Mattingly, I. G. (1985). The motor theory of speech perception revised.
Cognition, 21, 1–36
Meister, I. G., Wilson, S. M., Deblieck, C., Wu, A. D., & Iacoboni, M. (2007). The essential role of premotor cortex in speech perception. Current Biology, 17(19), 1692–1696.
Pulvermuller, F., Huss, M., Kherif, F., & Moscoso del Prado Martin, F. (2006). Motor cortex
Rogalsky, C., Love, T., Driscoll, D., Anderson, S. W., & Hickok, G. (2011). Are mirror neurons the basis of speech perception? Evidence from five cases with damage to the purported human mirror system. Neurocase, 17(2), 178–187.
Roon, K. & Gafos, A. (2014). Perceptuo-motor effects of response-distractor compatibility in speech: beyond phonemic identity. Psychonomic bulletin & review. 22.
Seifritz, E., Esposito, F., Hennel, F., Mustovic, H., Neuhoff, G. J., Bilecen, D., Tedeschi, G., Scheffler, K., Di Salle, F. (2002). Spatiotemporal pattern of neural processing in the human auditory cortex. Science, 297(5587), 1706-1708.
Watkins, K. E., Strafella, A. P., & Paus, T. (2003). Seeing and hearing speech excites the motor system involved in speech production. Neuropsychologia, 41(8), 989–994.
Watkins, K., & Paus, T. (2004). Modulation of motor excitability during speech perception:
The role of Broca’s area. Journal of Cognitive Neuroscience, 16(6), 978–987.
Wilson, S. M., Saygin, A. P., Sereno, M. I., & Iacoboni, M. (2004). Listening to speech activates motor areas involved in speech production. Nature Neuroscience, 7, 701–702.
