Data processing: speech signal processing – linguistics – language – Speech signal processing – Application
Reexamination Certificate
2001-08-28
2003-12-09
McFadden, Susan (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Application
Reexamination Certificate
active
06662162
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to an acoustic based method of identifying and rating motor speech deterioration and the underlying pathologies of the deterioration.
BACKGROUND OF THE INVENTION
Ataxia is a profound loss of muscular coordination which characterizes cerebellar-based pathology (Diener & Dichgans, 1992). A loss of coordination in ataxia may take the form of the loss of balance, the inability to walk heel to toe, nystagmus, difficulty alternating sequences of movements, dysarthria (Diener, & Dichgans, 1992), dysmetria and hypermetria (Hallet, Shahani, & Young, 1975 a, b).
Kinematic/electromyographic (EMG) studies of speech have investigated both normal muscle activation patterns (Harris, 1978; Tuller, Harris, & Kelso, 1981) and muscle patterns in ataxic dysarthric speakers (Ackermann, Hertrich, Daum, Scharf, & Spieker, 1997; (Ackermannn, Hertrich, & Scharf, 1995). Several investigations have also focused on the possible similarities or differences between limb movements and speech movements in ataxia. (Gentil, Devanne, Maton, & Brice, 1992; Salisachs, 1979; Akermann et al., 1995; Ostry, Keller, & Parush, 1983).
Early descriptions of dysarthria (Darley, Aronson & Brown, 1969a., b) were based solely on the perceptual judgments of several speech and voice characteristics. The 10 speech and voice characteristics of ataxic dysarthria (Darley et al., 1969a), typified by poor coordination of the articulators, were grouped into three clusters: 1. articulatory inaccuracy (imprecise consonants, irregular articulatory breakdown, and distorted vowels); 2. prosodic excess (excess and equal stress, prolonged phonemes, prolonged intervals, and slow rate); and 3. phonatory-prosodic insufficiency (monopitch, monoloudness, and harsh voice). However, perceptual descriptions are misleading and the distinctive patterns once claimed, do not hold for comparisons of neurological syndromes (Ziegler) or for those patterns found in ataxic speakers.
Subsequent studies attempted to supplement the perceptual description of ataxic dysarthria with acoustic analyses (see e.g., Kent, Netsell, & Abbs, 1979), however, they did not describe the prosodic descriptors in acoustic terms nor correlated their analyses with neurological pathological conditions.
Acoustic descriptions of ataxic speech were used to measure duration and first and second formant onset/offset frequencies. In studies using ataxic subjects, Kent et al. used acoustic measurement to describe ataxic speech and concluded that that formant frequencies were normal in ataxic subjects while duration measures were not. However the speaker tasks were many and did not capture the disturbance of syllables in several contexts. Thus whether this finding held for words within a phrase, or for syllables or words within sentences was not known. Further, although the acoustic analysis of dysarthria caused by cerebellar damage found a disproportionate lengthening of the segment to be a fundamental property of ataxic dysarthria (Kent et al., 1979), it was not known whether this was a property specific to ataxic dysarthria, or was present in all or some other dysarthrias. Examining narrow band spectrograms of sentences led Kent et al. to suggest a syllable-level planning with a falling f
0
on each successive syllable. The lengthening of segments and syllables led the investigators to posit a disordered prosody for cerebellar subjects.
Prosody in normal speech production has included descriptions of F
0
, formant frequencies and syllable duration. These acoustic descriptions have not described the dissociation between time (duration) and space (oral pharyngeal space) inferred from formant frequency values F
1
and F
2
. This dissociation is important because it allows for description of pathological utterances that are long and reduced in movements (slurred speech) and the slowing down of normal speech that occurs at the end of utterances. Previously models of speech predicted that all lengthened syllables would have more extreme movements.
A critical issue in the study of speech motor control is the identification of the mechanisms that generate the temporal flow of serially ordered articulatory events. Early investigations were motivated by Lashley's (1951) model that predicted a monotonic relationship between vowel duration and formant frequency. Lindblom (1963), for instance, claimed that articulatory “undershoot” is the basis for any reduction in vowel duration from normal values.
Subsequent studies did not find that duration and formant frequency were monotonically linked. Harris (1978) found that, contrary to the predictions of Lindblom's model, when either rate or stress was manipulated, syllable duration and vowel formant frequency varied independently in a non-monotonic relation. In addition, EMG studies showed reduced orbicularis oris and genioglossus activity for syllables of reduced stress (Tuller et al., 1981; Harris, 1971,1978). The conclusions drawn from these physiological and acoustic data of normal speakers was that any change in rate or stress may result in independent variations of syllable duration and formant values.
The components of prosody have been defined as the acoustic features of f
0
, segment duration, amplitude and segmental quality. Variations in the values of these features signal, among other things, constituent boundaries and syllable prominence. The kinematic data of Cohen et al. (1995) for six different conditions of syllable prominence show a difference in velocity for accented syllables in phrase final vs. non-final position. This kinematic finding speaks to the non-monotonic relation between duration and formant frequencies: all durations are not the same. Sometimes speakers slow down, resulting in reduced vowel space with longer durations; thus there is a dissociation between duration and formant frequency. However, this was not shown in acoustic measures.
It is an object of the invention to provide a reliable quantitative acoustic assessment to describe the speech and voice characteristics of subjects with neuro-motor speech disturbances. It is a further object of the invention to provide a method that can correlate prosodic descriptors in acoustic terms with neurological pathological conditions.
SUMMARY OF THE INVENTION
The present invention provides a method of identifying speech motor dysfunction in a test subject comprising measuring one or more acoustic parameters of one or more prosodic conditions; comparing each acoustic parameter between pairs of prosodic conditions to obtain a contrast value; and comparing the contrast values for each acoustic parameter to contrast values of a normal subject, wherein a difference in contrast values between the test subject and the normal subject is correlated to speech motor dysfunction.
The present invention also provides a method of identifying speech deterioration in a test subject comprising measuring one or more acoustic parameters of one or more prosodic conditions; comparing each acoustic parameter between pairs of prosodic conditions to obtain a contrast value; and comparing the contrast values for each acoustic parameter to contrast values of a normal subject, wherein a difference in contrast values between the test subject and the normal subject is correlated to speech deterioration.
The present invention further provides a method of diagnosing speech motor dysfunction in a test subject comprising measuring one or more acoustic parameters of one or more prosodic conditions; comparing each acoustic parameter between pairs of prosodic conditions to obtain a contrast value; and comparing the contrast values for each acoustic parameter to contrast values of a normal subject, wherein a difference in contrast values between the test subject and the normal subject is correlated to speech motor dysfunction.
In another embodiment, the present invention provides a method of rating the severity of speech motor dysfunction in a test subject comprising measuring one or more acoustic parameters of one or more prosodic conditions; comparing each ac
Darby & Darby
McFadden Susan
LandOfFree
Method of rating motor dysfunction by assessing speech prosody does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of rating motor dysfunction by assessing speech prosody, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of rating motor dysfunction by assessing speech prosody will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3160953