Detailed Protocol - MIT



DETAILED PROTOCOL

Constraints and Strategies in Speech Production

April 25July 23, 2003

The current proposal is for two fMRI studies that are included in a renewal application for an NIH-funded project from the Research Laboratory of Electronics, M.I.T., “Constraints and Strategies in Speech Production,”, 2 R01 DC01925 – Joseph S. Perkell, Ph.D., D.M.D., P.I. The studies are to be directed by Prof. Frank Guenther, (Dept. of Cognitive and Neural Systems, Boston University) as a consultant to the M.I.T. project, and they also form a logical extension of a project of Dr. Guenther’s that is currently underway at the MGH NMR Center, “Neural Modeling and Imaging of Speech.” Much of the following detailed protocol is taken from the Neural Modeling and Imaging protocol, and differs only in the specifics of the two new studies.

I. BACKGROUND AND SIGNIFICANCE

The speech research literature is at a stage where it is important to formulate models that address not just a single speech phenomenon, such as motor equivalence or coarticulation, in isolation, but instead treat these phenomena within a comprehensive theoretical framework that can account for many aspects of speech. Because the dynamics of such a model are necessarily complex and its properties typically difficult to clearly visualize, a means for objective verification of the model’s properties must also be available. To meet these requirements, it is important to formulate computational models that are described in sufficient mathematical detail that their properties can be verified through computer simulation.

The speech research literature currently contains few comprehensive computational models of this type. Over the past five years, our laboratory at Boston University has developed, tested, and refined a computational model that extends the current state of the field in several directions. This framework is one of the first and most comprehensive computational frameworks that investigates auditory spaces for planning speech movements (Guenther, Hampson, and Johnson, 1998). The use of self-organization in our framework allows incorporation of speaker-specific vocal tract models, thus providing a more accurate and reliable means of interpreting experimental data and generating predictions for future experiments (Nieto-Castanon and Guenther, 1999). Finally, our modeling framework is one of the first computational frameworks in which the development of speech perception and production are addressed. The current proposal describes several projects designed to further develop this framework (see Specific Aims section).

2) Models of speech perceptual and motor development.

An understanding of the development of speech skills is important for many reasons. Potential health benefits include better diagnosis of developmental difficulties in infants and better therapies and prosthetic devices for overcoming these difficulties. We believe our modeling framework is the most developed adaptive neural network treatment of speech perceptual and motor development in the current scientific literature.

As speakers, we are capable of amazingly fast, flexible, and efficient movements. Competencies such as motor equivalence and coarticulation greatly increase the efficiency of articulator movements. A model of motor skill acquisition should embody these competencies, suggesting that we should work toward an understanding of speech motor skill acquisition within a comprehensive computational model of speech production. Our modeling framework has been used to address several speech motor development issues (Guenther, 1995; Callan et al., in press). Our research into motor skill acquisition has also led to a better understanding of speech production in adults: a convex region theory of the targets of speech, developed to explain how infants learn language-specific and phoneme-specific limits on articulatory variability, provides a unified explanation for such long-studied speech phenomena as contextual variability, coarticulation, motor equivalence, and speaking rate effects (Guenther, 1995).

As listeners, our brains develop so as to allow very rapid and effortless parsing of a continuously changing acoustic signal into a discrete set of phonemes, syllables, and words from our native language. This process appears to be aided by the fact that our perceptual spaces are warped such that we are more sensitive to between-category differences than within-category differences, as evidenced by phenomena such as categorical perception and the perceptual magnet effect. We have developed, tested, and refined a self-organizing neural network model that addresses this issue (Guenther and Gjaja, 1996; Guenther et al., 1999b). This model explains the warping of auditory perceptual space in terms of the development of neural maps in the auditory system. This work includes an account of the formation of vowel categories through simple exposure to phonemes from an infant’s native language.

Further studies into developmental aspects of our modeling framework are proposed in the current research plan, including development of auditory categories and visual influences on speech perception (see Specific Aims section).

3) fMRI as a tool for studying brain processes.

Functional magnetic resonance imaging (fMRI) was first introduced as a tool for measuring brain activation by Belliveau et al. (1991). The type of functional magnetic resonance imaging (fMRI) that we plan to use takes advantage of changes in blood oxygen levels that result when a particular part of the brain is activated. It has long been known that neural activation leads to increased local blood flow (Roy and Sherrington, 1890). More recently, this relationship was shown to exist in the absence of a proportional increase in oxygen consumption (Fox and Raichle, 1986). This results in an overall higher concentration of blood oxygen in an activated area. The concentration of deoxyhemoglobin, a paramagnetic substance which can be detected by the magnetic resonance scanner, varies inversely with blood oxygen level. The oxygenation state of the blood thus serves as a natural contrast agent in the brain, and the scanner is able to determine the relative activation of different areas of the brain by measuring the relative concentrations of deoxyhemoglobin in those areas. This is typically referred to as the BOLD (blood oxygenation level-dependent) fMRI method (Kwong et al. 1992). This technique provides a completely non-invasive method for producing maps of brain activity. These maps can then be used to compare relative activation of different areas of the brain while a subject is asked to perform a task. Typically this is done by looking at changes in activation during an experimental task as compared to a carefully chosen control task.

In recent years, fMRI has provided a wealth of information regarding brain function. Because it is safe for use with human subjects, this technique is particularly suited to the study of speech and language. Earlier techniques for studying brain function were invasive, limiting their use to animal subjects, although a small amount of human data came from studies of epileptic patients prior to surgery (e.g., Penfield and Roberts, 1959; Creutzfeldt, Ojemann, and Lettich, 1989). Studies of the effects of brain lesions in aphasic patients also provided clues (see Goodglass, 1993, for a review), but these data are difficult to interpret due to the large variability in lesion sites across subjects, small numbers of subjects, and coarse “grain” of the resulting data. Thus, relatively little was known about brain function during speech and language tasks until the late 1980’s.

4) Neural models for interpretation of functional brain imaging data.

The relatively recent advent of functional brain imaging techniques that are safe for use on human subjects has led to an explosion in the amount of data concerning brain activity during speech and language tasks. Examples include fMRI studies (e.g., Binder et al., 1997; Calvert et al., 1997; Celsis et al., 1999; Friedman et al., 1998; Neville et al., 1998; Rueckert et al., 1994; Small et al., 1996; Wildgruber et al., 1996), positron emission tomography (PET) studies (e.g., Demonet et al., 1992, 1994; Fiez et al., 1995; Friston et al., 1991; Hirano et al., 1996, 1997; Mazoyer et al., 1993; McGuire et al., 1996; Petersen et al., 1988, 1989, 1990; Wise et al., 1991, 1999; Zatorre et al., 1992), electroencephalogram (EEG) studies (e.g., Martin-Loeches, Schweinberger, and Sommer, 1997; Mills, Coffey-Corina, and Neville, 1993; Neville et al., 1993; van Turennout, Hagoort, and Brown, 1998), and magnetoencephalogram (MEG) studies (e.g., Numminen and Curio, 1999; Numminen, Salmelin, and Hari, 1999; Salmelin et al., 1999; Sams et al., 1991; Szymanski, Rowley, and Roberts, 1999). As these data accumulate, it is becoming increasingly important to have a modeling framework within which to interpret data from the various studies. Without such a framework, these important data can seem like a random set of information points rather than a coherent description of the neural processes underlying speech.

As described above, our laboratory at Boston University has been developing a neural network modeling framework that explains a wide range of speech perception and production phenomena, including perceptual warping in the auditory system, contextual variability in speech production, motor equivalence, speaking rate effects, coarticulation, and the acquisition of speaking skills. This modeling framework has been used to interpret data from a range of behavioral studies (Guenther, 1995; Guenther and Gjaja, 1996; Guenther, Hampson and Johnson, 1998), and new experiments have been performed to test specific model predictions and refine the models when necessary (Guenther et al., 1999a; Guenther et al., 1999b).

Because our modeling framework is defined using adaptive neural networks, interpreting the model's components in terms of brain regions and activations is a relatively straightforward process. We feel that this fact, in combination with the framework's success in explaining psychophysical and behavioral data concerning speech perception and production, makes it ideally suited for interpreting the results of current speech-related imaging studies and guiding future imaging studies. In the current application, we propose two fMRI experiments that are designed to contribute to this objective.

By refining our neural modeling framework through experimental tests and modeling projects, we will move closer to an understanding of the neural processes underlying speech and language that we believe will prove useful for addressing various speech-related diseases and injuries. Our model has already been used to interpret data from deaf individuals who have been given cochlear implants (Perkell et al., in press). In the current proposal, we propose to further this connection to clinical issues in various ways, including studies of cerebellar involvement in the speech of normal subjects and ataxic dysarthrics and the effects of lesions in different brain regions.     

II. SPECIFIC AIMS

The primary goal of this research project is the continued development, testing, and refinement of a comprehensive computational modeling framework addressing the neural processes underlying speech perception and production. This framework is defined using adaptive neural networks, allowing comparisons with data from imaging studies of brain function. The aims of the current two studies is to investigate mechanisms of feedback and feedforward control of speech movements through the use of fMRI in combination with perturbations of somatosensory and auditory feedback. We expect these experiments to provide valuable data regarding the neural processes underlying speech production and perception. This data will be used to refine our neural model.

III. SUBJECT SELECTION

Inclusion criteria: healthy right-handed male and female adults between the ages of 18 and 50 will be recrutied for this study.

Exclusion criteria: left handedness, native language other than American English, history of serious head injury (with loss of conciousness), history or current diagnosis of neurological disorder, current injury to hands or arms that would impede task performance, pregnancy, MR incompatibility (see attached MR screening sheet); subjects with a history of epilepsy or other seizure disorder will be exluded from studies involving visual stimuli.

Subjects will be recruited via advertisements posted at the MGH and the M.I.T. and Boston University campuses, via electronic mails, advertisements in student newspapers, and via word of mouth.

IV. SUBJECT ENROLLMENT

We will enroll 34 subjects in a total of two fMRI experiments -- 17 subjects per experiment. Subjects will be screened over the telephone or in person to ensure that they meet the basic inclusion-exclusion criteria for the study. On the day of the study, just prior to scanning, subjects will read and sign the informed consent document detailing the general purposes and procedures of the experiment, and any questions the subject has will be answered.

Informed consent will be obtained prior to each experimental session by the Principal Investigator, Dr. Julie GoodmanFrank Guenther, or a study co-investigator. The informed consent clearly states that the subject may choose to terminate the experiment at any time.

V. STUDY PROCEDURES

Subjects will undergo telephone screening to ensure that they meet the basic inclusion/exclusion criteria for the study. On the day of the study, subjects will give informed consent, and will be screened again for MR compatibility. This will be followed immediately by the fMRI scanning session. Some subjects will be asked to perform psychophysical tests immediately prior to MRI scanning. These tests will consist of word stem completion tasks that will require subjects to complete a word presented on a computer monitor, either verbally, or by typing on a keyboard. The tests are expected to take approximately 1 hour to complete. Each scanning session starts with the subject lying on the table of the MRI scanner. Blankets and pillows are used to help insure the comfort of the subject, which is very important for preventing unwanted movements during the 2-hour scanning session. The subject’s head is immobilized using foam pads inserted between the subject’s head and the head carriage of the scanner. The table is then slid into a large magnet with a bore of approximately 1 meter. During the first fifteen minutes or so of scanning, a conventional high-resolution anatomical scan is obtained to allow localization of brain structures. After these localizing images are obtained, functional images are obtained using the high-speed function of the MRI scanner, which is capable of measuring changes in blood flow that correlate with brain function. High-speed images are collected during 5 to 10 experimental runs, each lasting approximately 4 to 6 minutes. During these runs, subjects will be asked to attend to auditory stimuli, attend to visual stimuli, produce speech, and/or lie silently in the scanner. Auditory stimuli will be presented via insert headphones (see below) at a volume deemed comfortable by the subject. If no auditory stimuli are used in an experiment, earplugs are inserted to protect the subject from the sounds created by the scanner during imaging. Visual stimuli are projected onto a screen that the subject views through a mirror. The flashing pattern displayed by the video projector does not present any health hazards to normal volunteers. Subjects with a history of epilepsy or other seizure disorder will be excluded from studies involving visual stimuli. Subjects with corrected-to-normal vision who participate in experiments with visual stimuli will be provided with non-magnetic corrective glasses, available at the MGH MRI facility. After the high-speed imaging runs are completed, a few additional conventional images are collected to help with registering the functional data with the anatomical data. The entire imaging session lasts approximately 2 hours in addition to any psychophysical testing.

One of the two fMRI experiments involves the use of a pneumatically operated device that we have developed for perturbing mandibular closing movements. The device consists of a small, tubular-shaped inelastic balloon that is held in place between the molar teeth on one side. The balloon is connected via stiff plastic tubing to a small air cylinder driven by a powerful solenoid. Activation of the solenoid causes inflation of the balloon to a diameter of about 1 cm at a pressure of 4-5 psi within 100 ms. Under control of the computer that displays stimuli to the subject and records acoustic and movement signals, the solenoid can be activated for selected utterances at a predetermined delay from the onset of the subject's voicing. To avoid giving the subject auditory cues about the perturbation, the solenoid and air cylinder are located in the MRI control room. The other fMRI experiment involves the use of a sensorimotor adaptation apparatus that we have developed to modify the acoustic structure of speech sounds that are fed back to the subject (with a very short delay) as he or she pronounces simple utterances. Insert earphones are used, along with a slightly amplified feedback signal to effectively prevent the subject from hearing his own air-conducted vowel formant structure. Both of these perturbation devices have been tested and used successfully and safely in the speech physiology laboratory in the Speech Communication Group, R.L.E., M.I.T. The possibility of these types of perturbation is addressed in the fMRI consent form.

If, during a scanning session, an abnormality is suspected, an immediate consultation by the radiologist on duty at Bay 1 (the clinical magnet) will be requested while the subject is still in the magnet. Additional images will be acquired if deemed necessary by the radiologist. If clinical follow-up is recommended, the radiologist, not the co-investigator conducting the research, will convey this to the subject. If no radiologist is available for immediate consultation, films will be printed of the brain area in question. If the potential abnormality is detected on the T1-weighted image, an additional T2-weighted scan will also be acquired. At the earliest opportunity following the research study, a consultation with Greg Sorensen, MD, or another available radiologist. Under no circumstances will the potential abnormality be discussed with the subject before a radiologist has been consulted and the situation has been explained to the subject by a medical professional. Subjects will be informed that if an abnormality is suspected, a radiologist will be consulted

VI. BIOSTATISTICAL ANALYSIS

The primary data variables we will analyze are voxel activation values. A control condition, such as lying quietly or passively observing a blank screen, is employed to serve as an activation baseline. T-tests are then used to determine which voxels have a significantly different level of activation in experimental conditions as compared to the control condition. Regions of interest (ROIs) will be identified using anatomical scans collected at the beginning of each scanning session. Multivariate analysis of variance (ANOVA) will be used to determine differences in the percentage of active voxels within an ROI across conditions, as well as to investigate laterality effects and differences between ROIs.

From past experience, we believe that approximately 15-17 subjects will be needed to reach statistical significance in the above-mentioned comparisons for each experiment. Keeping in mind that scanner data collection problems occur fairly frequently, we believe that good data from 15-17 subjects can be achieved with 15 scanning sessions per experiment.

The study will begin on 12/1/2003, and it will end after data collected from all subjects has been analyzed (by 11/30/2007).

VII. RISKS AND DISCOMFORTS

There are no known or foreseeable risks or side effects associated with conventional MRI procedures except for those individuals who have electrically, magnetically, or mechanically activated implants (such as cardiac pacemakers or cochlear implants) or those who have intracerebral vascular clips. There are no known additional risks associated with high-speed MRI. Both the conventional and high-speed MRI systems have been approved by the FDA and will be operated within the operating parameters reviewed and accepted by the FDA. There are no known risks or discomforts associated with either of the perturbation devices.

As described above, careful screening for contra-indicators is conducted before a subject is enrolled in the study. Just prior to entering the scanning suite, subjects will be orally interviewed again to ensure that they are metal free. Periodically throughout the duration of the MRI exam, the examiner will confirm with the subject that they are still comfortable and wish to continue. This "participant status check" will occur after each series of images is acquired, and any adjustments required to facilitate subject comfort will be made. All subjects will be provided with mandatory hearing protection in the form of earplugs, or headphones, for experiments involving auditory stimuli, which will prevent discomfort due to scanner noise.

VIII. POTENTIAL BENEFITS

Subjects derive no benefit from the procedures except payment. However, risk is negligible and their participation will contribute very useful information concerning the mechanisms of speech perception and production. Thus the risk/benefit ratio is negligible. Subjects will be paid $100 for an experimental session lasting approximately 2 hours plus $10/hour for any testing done immediately prior to scanning..

IX. MONITORING AND QUALITY ASSURANCE

The technical staff of the MGH NMR Center will continually monitor the nature and quality of the data collection on the 3T magnet system. Modifications to enhance data quality and statistical power will be performed when appropriate.

X. REFERENCES

Belliveau, J.W., Kennedy, D.N., McKinstry, R.C., Buchbinder, B.R., Weisskoff, R.M., Cohen, M.S., Vevea, J.M., Brady, T.J., Rosen, B.R. (1991). Functional mapping of the human visual cortex by magnetic resonance imaging. Science, 254, pp. 716-719.

Binder, J.R., Frost, J.A., Hammeke, T.A., Cox, R.W., Rao, S.M., and Prieto, T. (1997). Human brain language areas identified by functional magnetic resonance imaging. Journal of Neuroscience, 17, pp. 353-362.

Callan, D.E., Kent, R.D., Guenther, F.H., and Vorperian, H.K. (in press). An auditory-feedback-based neural network model of speech production that is robust to developmental changes in the size and shape of the articulatory system. Journal of Speech, Language, and Hearing Research, in press.

Calvert, G.A., Bullmore, E.T., Brammer, M.J., Campbell, R., Williams, S.C.R., McGuire, P.K., Woodruff, P.W.R., Iversen, S.D., and David, A.S. (1997). Activation of auditory cortex during silent lipreading. Science, 276, pp. 593-596.

Celsis, P., Boulanouar, K., Ranjeva, J.P., Berry, I., Nespoulous, J.L., and Chollet, F. (1999). Differential fMRI responses in the left posterior superior temporal gyrus and left supramarginal gyrus to habituation and change detection in syllables and tones. NeuroImage, 9, pp. 135-144.

Creutzfeldt, O., Ojemann, G., and Lettich, E. (1989). Neuronal activity in the human lateral temporal lobe. II. Responses to the subjects own voice. Experimental Brain Research, 77, pp. 476-489.

Demonet, J.-F., Chollet, F., Ramsay, S., Cardebat, D., Nespoulous, J.-L., Wise, R., Rascol, A., and Frackowiak, R. (1992). The anatomy of phonological and semantic processing in normal subjects. Brain, 115, pp. 1753-1768.

Demonet, J.-F., Price, C., Wise, R., Frackowiak, R.S.J. (1994). Differential activation of right and left posterior sylvian regions by semantic and phonological tasks: A positron-emission tomography study in normal human subjects. Neuroscience Letters, 182, pp. 25-28.

Fiez, J.A., Raichle, M.E., Miezin, F.M., and Petersen, S.E. (1995). PET studies of auditory and phonological processing: Effects of stimulus characteristics and task demands. Journal of Cognitive Neuroscience, 7, pp. 357-375.

Fox, P.T. and Raichle, M.E. (1986) Focal physiological uncoupling of cerebral blood flow and oxidative metabolism during somatosensory stimulation in human subjects. Proceedings of the National Academy of Sciences, 83, pp. 1140-1144.

Friedman, L., Kenny, J.T., Wise, A.L., Wu, D., Stuve, T.A., Miller, D.A., Jesberger, J.A., and Lewin, J.S. (1998). Brain activation during silent word generation evaluated with functional MRI. Brain and Language, 64, pp. 231-256.

Friston, K.J., Frith, C.D., Liddle, P.F., and Frackowiak, R.S. (1991). Investigating a network model of word generation with positron emission tomography. Proceedings of the Royal Society of London. Series B: Biological Sciences, 244, pp. 101-106.

Goodglass, H. (1993). Understanding aphasia. San Diego, CA: Academic Press.

Guenther, F.H. (1995). Speech sound acquisition, coarticulation, and rate effects in a neural network model of speech production. Psychological Review, 102, pp. 594-621.

Guenther, F.H., and Gjaja, M.N. (1996). The perceptual magnet effect as an emergent property of neural map formation. Journal of the Acoustical Society of America, 100, pp. 1111-1121.

Guenther, F.H., Hampson, M., and Johnson, D. (1998). A theoretical investigation of reference frames for the planning of speech movements. Psychological Review, 105, pp. 611-633.

Guenther, F.H., Espy-Wilson, C.Y., Boyce, S.E., Matthies, M.L., Zandipour, M., and Perkell, J.S. (1999a). Articulatory tradeoffs reduce acoustic variability during American English /r/ production. Journal of the Acoustical Society of America, 105, pp. 2854-2865.

Guenther, F.H., Husain, F.T., Cohen, M.A., and Shinn-Cunningham, B.G. (1999b). Effects of categorization and discrimination training on auditory perceptual space. Journal of the Acoustical Society of America, 106, pp. 2900-2912.

Hirano, S., Kojima, H., Naito, Y., Honjo, I., Kamoto, Y., Okazawa, H., Ishizu, K., Yonekura, Y., Nagahama, Y., Fukuyama, H., and Konishi, J. (1996). Cortical speech processing mechanisms while vocalizing visually presented languages. NeuroReport, 8, pp. 363-367.

Hirano, S., Kojima, H., Naito, Y., Honjo, I., Kamoto, Y., Okazawa, H., Ishizu, K., Yonekura, Y., Nagahama, Y., Fukuyama, H., and Konishi, J. (1997). Cortical processing mechanism for vocalization with auditory verbal feedback. NeuroReport, 8, pp. 2379-2382.

Kwong, K.K., Belliveau, J.W., Chesler, D.A., Goldberg, I.E., Weisskoff, R.M., Poncelet, B.P, Kennedy, D.N., Hoppel, B.E., Cohen, M.S., Turner, R., Cheng, H, Brady, T.J., Rosen, B.R. (1992). Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. Proceedings of the National Academy of Sciences, 89, pp. 5675-5679.

Martin-Loeches, M., Schwienberger, S.R., and Sommer, W. (1997). The phonological loop model of working memory: An ERP study of irrelevant speech and phonological similarity effects. Memory & Cognition, 25, pp. 471-483.

Mazoyer, B.M, Tzourio, N., Frak, V., Syrota, A., Murayama, N., Levrier, O., Salamon, G., Dehaene, S., Cohen, L., and Mehler, J. (1993). The cortical representation of speech. Journal of Cognitive Neuroscience, 5, pp. 467-479.

McGuire, P.K., Silbersweig, D.A., Murray, R.M., David, A.S., Frackowiak, R.S.J., and Frith, C.D. (1996). Functional anatomy of inner speech and auditory verbal imagery. Psychological Medicine, 26, pp. 29-38.

Mills, D.L., Coffey-Corina, S.A., and Neville, H.J. (1993). Language acquisition and cerebral specialization in 20-month-old infants. Journal of Cognitive Neuroscience, 5, pp. 317-334.

Nieto-Castanon, A., and Guenther, F.H. (1999). Constructing speaker-specific articulatory vocal tract models for testing speech motor control hypotheses. Proceedings of the XIVth International Congress of Phonetic Sciences, pp. 2271-2274. Berkeley: Regents of the University of California.

Neville, H.J., Coffey, S.A., Holcomb, P.J., and Tallal, P. (1993). The neurobiology of sensory and language processing in language-impaired children. Journal of Cognitive Neuroscience, 5, pp. 235-253.

Neville, H.J., Bavelier, D., Corina, D., Rauschecker, J., Karni, A., Lalwani, A., Braun, A., Clark, V., Jezzard, P., and Turner, R. (1998). Cerebral organization for language in deaf and hearing subjects: Biological constraints and effects of experience. Proceedings of the National Academy of Sciences, 95, pp. 922-929.

Numminen, J., and Curio, G. (1999). Differential effects of overt, covert, and replayed speech on vowel-evoked responses of the human auditory cortex. Neuroscience Letters, 272, pp. 29-32.

Numminen, J., Salmelin, R., and Hari, R. (1999). Subject’s own speech reduces reactivity of the human auditory cortex. Neuroscience Letters, 265, pp. 119-122.

Penfield, W., and Roberts, L. (1959). Speech and brain mechanisms. Princeton, NJ: Princeton University Press.

Perkell, J.S., Guenther, F.H., Lane, H., Matthies, M.L., Perrier, P., Vick, J., Wilhelms-Tricarico, R., and Zandipour, M. (in press). A theory of speech motor control and supporting data from speakers with normal hearing and profound hearing loss. Journal of Phonetics, in press.

Petersen, S.E., Fox, P.T., Posner, M.I., Mintun, M.A., and Raichle, M.E. (1988). Positron emission tomographic studies of the cortical anatomy of single-word processing. Nature, 331, pp. 585-589.

Petersen, S.E., Fox, P.T., Posner, M.I., Mintun, M.A., and Raichle, M.E. (1989). Positron emission tomographic studies of the processing of single words. Journal of Cognitive Neuroscience, 1, pp. 153-170.

Petersen, S.E., Fox, P.T., Snyder, A.Z., and Raichle, M.E. (1990). Activation of extrastriate and frontal cortical areas by visual words and word-like stimuli. Science, 249, pp. 1041-1044.

Roy, C.W., and Sherrington, C.S. (1890). On the regulation of blood supply of the brain. Journal of Physiology, 11, pp. 85-108.

Rueckert, L., Appollonio, I., Grafman, J., Jezzard, P., Johnson, R. Jr., Le Bihan, D., and Turner, R. (1994). Magnetic resonance imaging functional activation of left frontal cortex during covert word production. Journal of Neuroimaging, 4, pp. 67-70.

Salmalin, R., Schnitzler, A., Parkkonen, L., Biermann, K., Helenius, P., Kiviniemi, K., Kuukka, K., Schmitz, F., and Freund, H. (1999). Native language, gender, and functional organization of auditory cortex. Proceedings of the National Academy of Sciences, 96, pp. 10460-10465.

Sams, M., Aulanko, R., Hamalainen, M., Hari, R., Lounasmaa, O.V., Lu, S.T., and Simola, J. (1991). Seeing speech: Visual information from lip movements modifies activity in the human auditory cortex. Neuroscience Letters, 127, pp. 141-145.

Small, S.L., Noll, D.C., Perfetti, C.A., Hlustik, P., Wellington, R., and Schneider, W. (1996). Localizing the lexicon for reading aloud: Replication of a PET study using fMRI. NeuroReport, 7, pp. 961-965.

Szymanski, M.D., Rowley, H.A., and Roberts, T.P. (1999). A hemispherically asymmetrical MEG response to vowels. NeuroReport, 10, p. 2481-2486.

van Turennout, M., Hagoort, P., and Brown, C.M. (1998). Brain activity during speaking: From syntax to phonology in 40 milliseconds. Science, 280, pp. 572-574.

Wildgruber, D., Ackermann, H., Klose, U., Kardatzki, B., and Grodd, W. (1996). Functional lateralization of speech production: A fMRI study. NeuroReport, 7, pp. 2791-2795.

Wise, R., Chollet, F., Hadar, U., Friston, K., Hoffner, E., and Frackowiak, R. (1991). Distribution of cortical neural networks involved in word comprehension and word retrieval. Brain, 114, pp. 1803-1817.

Wise, R.J., Greene, J., Buchel, C., and Scott, S.K. (1999). Brain regions involved in articulation. Lancet, 353, pp. 1057-1061.

Zattore, R.J., Evans, A.C., Meyer, E., and Gjedde, A. (1992). Lateralization of phonetic and pitch discrimination in speech processing. Science, 256, pp. 846-849.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download