Connected Discourse Tracking Research Articles

Evaluations of single-channel and multichannel tactile aids for the hearing impaired

A number of single-channel and multichannel tactile devices for hearing-impaired persons have been evaluated in our laboratory over the last several years. Both commercially available and experimental devices have been tested, including the Tactaid I, II, and V, Minifonator, Minivib, Queen's University tactile vocoder (both laboratory and wearable versions), and Tacticon TC-1600. Normal-hearing and hearing-impaired adults, and hearing-impaired children, have participated in a variety of training tasks, employing both recorded and live-voice stimuli. These tasks include simple detection of sound, environmental sound identification, syllable rhythm and stress categorization, phoneme identification, word identification, phrase and sentence identification, connected discourse tracking, and a question-and-answer “conversation” task. In the present paper, the results from these evaluations are considered as a whole, to permit generalizations about the kinds of information that can be provided by single-channel, dual-channel, and multichannel devices, with and without the addition of lipreading. In addition, results are discussed in terms of the development of “optimal” training procedures for the use of tactile aids. [Work supported by NSF and NIH.]

Evaluation of a multichannel electrotactile device for the hearing impaired—A case study

A multichannel electrotactile device for the hearing impaired was evaluated as an aid for speech reception and production with a 12‐year‐old profoundly hearing‐impaired child enrolled in an oral school for the deaf. Over an 8‐month period the subject received training in speech reception skills in a hierarchy of tasks ranging from discrimination of minimal paired words to connected discourse tracking. Aspects of the subject's speech production skills were evaluated by obtaining ratings from a group of teachers of the hearing impaired, who viewed a videotape of the child during syllable production and connected discourse tracking with and without the device. Results of speech reception testing indicated that the device allowed good discrimination of minimal paired words based on manner contrasts, but poor discrimination based on place contrasts. During tracking, lipreading‐pulse‐tactile aid conditions were superior to lipreading alone. Results of the speech production evaluation showed that syllable identification was better and ratings on several aspects of speech production were higher under tactile aid conditions. [Work supported by NIH.]

Comparison of two multichannel tactile aids for the hearing-impaired

Two multichannel tactile devices for the hearing-impaired were compared in tasks using acoustic stimuli of varying levels of complexity. One device was a 16-element linear vibratory array worn on the forearm, which displayed activity in 16 overlapping frequency channels. The other device utilized 16 channels in a similar stimulus processing strategy, but delivered the tactile stimulation to a linear electrocutaneous array worn on the abdomen. Subjects were tested in tasks including phoneme discrimination in pairs of rhyming syllables, phoneme identification in larger sets of rhyming syllables, learning of a list of words, and connected discourse tracking. All stimuli were presented live-voice by a female talker, and lipreading was permitted only in the tracking task. Results showed both devices to be good transmitters of manner and voicing features of articulation, but poor transmitters of place features. Differences between the two devices were found only at the level of connected discourse tracking. Results are discussed in terms of differences between the two devices in processing strategy, location of stimulator array, and type of transducer. [Work supported by NSF and NIH.]

Use of connected discourse tracking to train functional speech skills.

The connected discourse tracking procedure was adopted for use as a production training task for profoundly hearing-impaired children. A clinician served as the receiver and 5 children with profound hearing losses each served as the sender for 16 tracking sessions. There were individual differences among the children in their ability to convey the connected discourse to the clinician over time. The increase in comprehension of each subject's speech during the last session relative to the amount of information understood during the first session ranged from 0 to 48%. Two independent observers were used during the first and last sessions to verify the tracking results obtained by the clinician (the primary receiver). There were differences in words tracked per minute among the three receivers with the primary clinician achieving higher tracking rates than the other two. The children learned how to use discourse strategies when they were not understood and they were highly motivated by the task. The results indicate that it is feasible to incorporate the procedure in a speech training program for hearing-impaired children.

Effect of Text Complexity upon Connected Discourse Tracking

Effect of Text Complexity upon Connected Discourse Tracking

The transmission of prosodic information via an electrotactile speechreading aid.

Variations in voice fundamental frequency were extracted from naturally produced speech samples and transmitted to an electrocutaneous display consisting of 10 electrodes arranged in a linear array along the forearm. Changes in fundamental frequency were encoded as changes in stimulus location. Speechreading performance, with and without the electrocutaneous aid, was examined for both sentence and connected discourse materials with one profoundly hearing-impaired listener and one normally hearing listener whose hearing was masked. After 20 hours of speechreading training in connected discourse tracking procedure, both subjects showed higher tracking rates and faster learning rates with the aid than without the aid. Results from closed-set sentence-identification tests showed that patterns of intonation, stress, and phrase structure, which are not easily speechread, are readily available with the aid.

The contribution of fundamental frequency, amplitude envelope, and voicing duration cues to speechreading in normal-hearing subjects.

The ability to combine speechreading (i.e., lipreading) with prosodic information extracted from the low-frequency regions of speech was evaluated with three normally hearing subjects. The subjects were tested in a connected discourse tracking procedure which measures the rate at which spoken text can be repeated back without any errors. Receptive conditions included speechreading alone (SA), speechreading plus amplitude envelope cues (AM), speechreading plus fundamental frequency cues (FM), and speechreading plus intensity-modulated fundamental frequency cues (AM + FM). In a second experiment, one subject was further tested in a speechreading plus voicing duration cue condition (DUR). Speechreading performance was best in the AM + FM condition (83.6 words per minute,) and worst in the SA condition (41.1 words per minute). Tracking levels in the AM, FM, and DUR conditions were 73.7, 73.6, and 65.4 words per minute, respectively. The average tracking rate obtained when subjects were allowed to listen to the talker's normal (unfiltered) speech (NS condition) was 108.3 words per minute. These results demonstrate that speechreaders can use information related to the rhythm, stress, and intonation patterns of speech to improve their speechreading performance.

Validation of the Audiometer Telephone Interface (ATI)

Speech discrimination scores and connected discourse tracking (CDT) measurements were used to compare the clinical use of the Audiometer Telephone Interface (ATI) [R. G. Stoker, Hear. Instrum. 33(6), 26–29, 46 (1982)] with actual telephone performance on a local telephone company loop. Thirty hearing‐impaired subjects from three hearing‐loss categories—moderate, sloping; moderate, precipitous; and severe, sloping were used. Subjects were evaluated using standard audiometric procedures (under headphones) and with the ATI (telephone handset with simulated telephone network). Testing was also performed at the subjects' home by providing each subject with an amplifier handset so that the test could be performed at a Most Preferred Listening level (MCL). Stimuli for both the clinical trial and the home trial were the nu‐6 word lists and connected discourse tracking using a novel passage with neutral emotional content. Results suggest that the ATI effectively simulates the average hearing difficulty experienced by hearing‐impaired telephone users when using their own phones over a local loop served by a variety of central office systems.