Frenchay Dysarthria Assessment What s new Rebecca Palmer Pam

Frenchay Dysarthria Assessment: What’s new? Rebecca Palmer, Pam Enderby, James Carmichael

Topics § Original FDA overview § Advantages and disadvantages of this assessment § FDA 2 – new aspects § Computerised FDA § Demonstration § Current work on automated intelligibility testing

Original FDA § Author: Pam Enderby § First published in 1983 § Result of research identifying nature and patterns of oromotor movements associated with different neurological diseases (Enderby 1983) § Translated into French, German, Dutch, Norwegian, Swedish, Finnish, Catalan and Castilian

Aim of FDA § To analyse several important parameters of the motor speech system § To guide treatment § To assist with neurological diagnosis § To have good reliability and validity between and within clinicians without extensive training

Structure of FDA § Reflexes § Cough, swallow, dribble/drool § Respiration § Lips § At rest, in speech § At rest, spread, seal, alternate, in speech § Palate § Fluids, maintenance, in speech § Laryngeal § Time, pitch, volume, in speech § Tongue § At rest, protrusion, elevation, lateral, alternate, in speech § Intelligibility § Words, sentences, conversation

Procedure § Ask patient to carry out a task § Rate ability of each parameter using a 9 point scale – 5 descriptors + ½ marks

Advantages of FDA § Intelligibility commonly used to assess severity of dysarthria and to monitor progress BUT Intelligibility measures alone do not diagnose type of dysarthria or guide treatment § FDA breaks speech up into its component parts so the clinician can analyse what contributes to the reduced intelligibility thus guiding treatment § FDA provides a profile that contributes to the neurological diagnosis

Disadvantages of FDA § Some measures can be subjective § Some descriptors are interpreted differently by different clinicians reducing reliability § Intelligibility section: § Too few words/sentences regular users can learn them § Sentence structure = ‘the man is…’ therefore only listening for the last word § Scoring system based on number listener understood out of 10 (crude)

FDA 2 § Authors: Pam Enderby & Rebecca Palmer § 2008 § Aim: To address theoretical and practical issues identified in reviews of the first edition

Improvements 1 § Omitted items that have been found to be unreliable or redundant to the purposes of diagnosis and treatment § e. g. Jaw tests – patients rarely have abnormality in the jaw therefore the information didn’t assist diagnosis

Improvements 2 § Improved reliability of descriptors § Inter-rater reliability testing between experienced users of the FDA showed that some descriptors were interpreted differently. § a) e) E. g. voice time Patient can say ‘ah’ for 15 seconds Patient unable to sustain clear voice for 3 seconds Constant hoarse voice – RP = a), PE = e)

Improvements 2 § Inter rater and test reliability § Audio recordings of 9 people with a range of types and severities of dysarthria performing the audible FDA 2 tests: § 6 speech therapists working with a mixed adult caseload judged 42 examples of FDA 2 tests. § Scored on a 9 point scale § Same 42 tests presented again to the listeners after 6 week interval § Inter and intra rater reliability were calculated using intra class correlation coefficients

Inter and intra judge reliability Judge 1 1 2 3 4 5 2 0. 77 0. 92 3 0. 56 0. 65 0. 72 4 5 0. 67 0. 38 0. 60 0. 52 0. 51 0. 49 0. 79 6 0. 66 0. 72 0. 73 0. 70 0. 49 0. 56 6 0. 76 Criteria for interpretation of reliability coefficients for ordinal measures (Landis & Koch, 1977): <0 = poor, 0. 01 -0. 20 = slight, 0. 21 -0. 40 = fair, 0. 41 -0. 60 = moderate (mod), 0. 61 -0. 80 = substantial (sub) 0. 81 – 1 = almost perfect (per)

Improvements 3 § In speech tests § Sound saturated sentences provided for patient to say so that clinician can listen to the accuracy of sound placement in speech Lips in speech: ‘Mary brought me a piece of maple syrup pie’ Tongue in speech: ‘Kenneth’s dog took ten tiny ducks today’

Improvements 4 § Intelligibility testing § New set of words § Corpus of 116 words to reduce probability of listeners learning the words with increased exposure § Phonetically balanced list for types of sounds, position of sounds in words, word length § Word frequency >10 per million to control for any effects of word frequency on intelligibility

Improvements 4 § Sentence intelligibility § Key words phonetically balanced to account for place, manner, position and word length § Carrier phrases/sentences are all different so the listener has to listen to a sentence, not just interpret the key word in a standard carrier phrase § § § ‘Can you go the shop? ’ ‘My daughter is a nurse’ ‘Lets go to theatre’

Availability § FDA 2 available now from Pro-ed § Only in English!

Computerised FDA § James Carmichael produced computer version § Demonstration

Planned additions to CFDA Automation of intelligibility testing – modelling the naiive listener § If the learning effect alters a listener’s perception of a particular individual’s speaking style, is that listener’s judgement still representative of the naïve listener? § Can a computer model be built which behaves like an “eternal” naïve listener (i. e. never adapting to an unfamiliar speaking style and therefore always consistent in assessment)?

Using HMM Models to Emulate the Naïve listener • A hidden Markov Model (HMM) • a statistical representation of a speech unit at the phone/word/utterance level. • HMM models are “trained” by analysing the acoustic features of multiple utterances representing the specified speech unit. Multiple Speech Samples from multiple speakers

Goodness of fit § Once trained, an HMM word model can be used to estimate the likelihood that a given speech sound could have actually been produced by that word model. § This likelihood is called a goodness of fit (GOF) § expressed as a log likelihood, e. g. 10 -35 (or simply expressed as -35).

Comparing GOF scores with Subjective Assessments § 3 important cues of intelligibility are: § hesitation time; § speech rate § a phoneme-by-phoneme comparison of what the speaker intended to say and what the listener actually heard.

Calculating Phonetic Convergence Phoneme comparison of intended and perceived message: “You have to pay” (for a mildly dysarthric speaker) Intended Heard /j/ /u: / /h/ /æ/ /v/ /t/ /u: / /p/ /e/ /j/ /u: / /h/ /æ/ /v/ /d/ /u: / /b/ /aι/ 1 1 0 1 0 0 Convergence 1 Word Level Deletion Overall Convergence -1 5 out of a possible 9 = 0. 56 (56%)

Phonetic convergence L 1 L L 1 Mild, Moderate, Severe Listener 5 5 0 s Speech rate Mild, Moderate, Severe Hesitation L 2 0 L 10 L 5 L 15 Mild, Moderate, Severe Listener s Speech rate’s correlation with intelligibility is not as good as hesitation time or phonetic convergence, so we derive a Perceptual Intelligibility Index (PII) based on the Phonetic Convergence score weighted by a hesitation time coefficient L 20

How well do automated GOF scores correlate with Perceptual intelligibility index? Speaker Phon. Convergence Hesitation Time coefficient Sentence PII Score Avg. GOF Score Mild 0. 95 0. 91 0. 86 -34 Moderate 0. 27 0. 15 0. 11 -61 Severe 0. 20 0. 19 0. 04 -85 Correlation between GOF scores and PII scores =0. 72 Automated scores of goodness of fit measures generated by HMMs could be a valid and consistent intelligibility measure

Summary § FDA 2 § Analyses each parameter of speech § Enables clinician to find cause of reduced intelligibility, § § § guiding treatment Assists with diagnosis of dysarthria type and neurological impairment Excludes redundant tests Uses non-ambiguous descriptors Has inter and intra-rater reliability Large corpus of words and sentences controlled for linguistic and phonetic parameters for intelligibility sections Word and sentence cards provided

Summary § Computerised FDA § Provides training test for new users § Automatically produces profile and stores information § Increases objectivity of measures § Provides visual feedback of performance and improvements to patient § Seeks to automate measurement of intelligibility leading to increased consistency

Thank you !