PubMed Health⌕ Search

Biomedical subjects

C L De Filippo

Publications and source records attributed to C L De Filippo.

7 recordsLinked to original sources
PubMed1998

Round-off error, blind faith, and the powers that be: a caution on numerical error in coefficients for polynomial curves fit to psychophysical data.

Graphing and statistics software often permits users to fit polynomial curves, like a parabola or sigmoid, to scatter plots of psychophysical data points. These programs typically calculate the curve using double- or extended-precision numerical algorithms and display the resulting curve overlaid graphically on the scatter plot, but they may simultaneously display the equation that generates that curve with numerical coefficients that have been rounded off to only a few decimal places. If this equation is used for experimental or clinical applications, the round-off error, especially on coefficients for the higher powers, can produce anomalous findings due to systematic and extreme distortions of the fitted curve, even artifactually reversing the algebraic sign of the true slope of the fitted curve at particular data points. Care must be exercised in setting round-off criteria for coefficients of polynomial terms in curve-fit equations to avoid nonsensical measurement and prediction.

Computer Graphics↗
PubMed1994

Deriving passage difficulties for a tracking study via the Cloze technique.

The current report demonstrates the importance of formally accounting for passage difficulty when using the tracking procedure. Cloze responses to 82 encyclopedia excerpts (343-349 words each) were obtained from a large pool of normal-hearing adults and scored verbatim. Passage difficulty, derived via ANOVA, was then defined as the deviation of a passage's mean Cloze score from the score for all passages, corrected for differences among respondents. The passage difficulties were applied in an alternating conditions tracking experiment with one adult cochlear implant user. Conditions included conventional auditory-visual and auditory-only tracking and experimental mode-switching techniques in which the talker changed modalities during the correction phase. An ANCOVA of the word-per-minute scores was conducted, with passage difficulty as a covariate and passage adjustment values as the output. Tracking rates and percentage of words correct from the beginning and end of training were examined. Use of adjusted data reversed the interpretation of performance change, demonstrating the need for determining passage difficulties a priori.

Audiometry↗
PubMed1991

Loudness judgment procedures for evaluating hearing aid preselection decisions for severely and profoundly hearing-impaired listeners.

Hearing aid fitting involves a two-phase process of preselection and evaluation (Seewald RC and Ross M. Amplification for the Hearing Impaired 1988:213-271). The purpose of the present study was to examine alternative procedures that clinicians might use in the evaluation phase to verify the adequacy of hearing aid preselection decisions for severely and profoundly hearing-impaired listeners. Bekesy tracking, loudness rating, and conventional bracketing procedures were used to determine threshold, most comfortable listening level, and uncomfortable listening level for 10 hearing-impaired young adults. Stimuli were pulsed pure tones of 500, 1000, and 2000 Hz and filtered words. Means and standard deviations of most comfortable listening levels and uncomfortable listening levels derived from loudness judgments of the 10 subjects showed only nominal differences across procedures. However, correlation analysis (Pearson r) indicated that individuals responded to the three procedures in varying ways, producing different loudness judgments and overall dynamic ranges. Thus, test procedure may influence the clinician's final evaluation of a preselected hearing aid. Initial work suggests that closed-set response categories such as loudness rating can limit measurement variability and potentially guide the clinician's evaluation of hearing aid preselection decisions.

Adult↗
PubMed1978

A method for training and evaluating the reception of ongoing speech.

The "tracking procedure" is a method to train and evaluate the reception of ongoing speech. A talker and a receiver engage in a dialogue for a designated period of time in which the receiver reports his perception of successive segments of read text and is corrected by the talker until the text is repeated verbatim. Performance is measured in number of words of text repeated correctly per unit of time. The procedure is illustrated by its application to a project for training and evaluating use of a vibrotacile-electrotactile aid to lipreading. Differences between experimental conditions are described over time in terms of absolute words per minute, increment of aided over unaided words-per-minute scores, and percent of normal listening rate. Data from identification tests with syllables, words, and sentences are also presented. Tracking is suggested as an additional procedure for communication training in aural rehabilitation or classroom setting, and for evaluation of other communication systems.

Auditory Perception↗
PubMedDate not supplied

Laboratory projects in tactile aids to lipreading.

Four studies are reported on the design of tactile aids to lipreading conducted at Central Institute for the Deaf, St. Louis. A speech feature aid represented resonance, envelope, and source characteristics of speech on three vibrators at the fingertips. Artificially-deafened and congenitally deaf adults lipreading with the aid obtained an average of 40% gain over lipreading alone on syllable identification tests, but minimal gain for untrained words. In a series of three other studies, the quality of tactile sensation was developed as a cue to spectral information, presented in a temporal code based on the overall time-intensity waveform of speech. Electrical and vibratory stimulation on the hand and torso were evaluated by artificially-deafened adults using the tracking technique with connected speech. Short-term training yielded variable increments over unaided lipreading scores and lipreading with a simple vibrator aid. Neither the feature nor the spectral code was superior in aiding lipreading. The displays that produced quality changes seemed more efficient for training than the spatial display. Results were affected also by content and method of training and initial lipreading proficiency.

Adolescent↗