MONAURAL LOW-REDUNDANCY SPEECH TESTS


Monaural presentation involves the presentation of one message to a single ear at a time. In monaural speech tests, the extrinsic redundancy is reduced by filtering and by time alteration such as compression or interruption.

A. LOW-PASS FILTERED
Bocca, Calearo, and Cassinari (1954) and Bocca, Calearo., Cassinari, and Migliavacca (1955) obtained performance-intensity functions for low-pass filtered (cutoff frequency of 500 Hz) lists of PB words in persons with temporal-lobe tumors with normal-hearing threshold levels. They found a contralateral-ear effect (i.e., scores were poorer in the ear contralateral to the side of the temporal-lobe lesion) that was absent (or reduced) when the lists were presented with­out frequency distortion. Similar findings were obtained by Goldstein etal. (1956), using the Rush Hughes poor-quality (presumably frequency distorted) recordings of PB words, and Jerger (1960a,b, 1964), using low-pass filtered PAL PB-50 word lists.


Lynn and Gilroy (1972, 1975, 1976, and 1977) have con­ducted several investigations on the effects of various cen­tral auditory lesions on low-pass filtered speech scores. For example, Lynn and Gilroy (1977) obtained the scores for low-pass filtered NU-6 monosyllabic PB words (cutoff-frequency of 500 Hz with a rejection rate of 34 dB/octave) presented at 60 dB SL relative to the PTA or SRT in a group of 11 patients with surgically confirmed right temporal-lobe tumors and 11 patients with left temporal-lobe tumors. The contralateral-ear effect was present in the right and left temporal-lobe tumor cases, with an average interaural dif­ference of approximately 16% for the right temporal-lobe cases and approximately 12% for the left temporal-lobe cases. Lynn and Gilroy (1977) found that, in their group of 35 patients with temporal-lobe tumors, 74% had contra­lateral-ear effects, 3% had ipsilateral-ear effects, and 24% had normal scores on the low-pass filtered speech test.

Studies (Lynn & Gilroy, 1977- Musiek, Wilson, & Pin-heiro, 1979) have revealed that the scores for both ears for low-pass filtered speech are generally unaffected by lesions of the interhemispheric pathways (corpus callosum). In a recent study, Baran, Musiek, and Reeves (1986) reported that post-surgical scores on the low-pass filtered speech test did not differ significantly from the preoperative scores in a group of 8 patients who underwent anterior section of the corpus callosum. Lynn and Gilroy (1977) found that, in a group of 38 patients with parietal-lobe tumors affecting the interhemispheric auditory pathways, 74% had normal scores, 22% had abnormal scores in the left ear, and 4% had abnormal scores in the right ear.

Consistent findings on low-pass filtered speech tests have not been obtained for brainstem lesions. Ipsilateral, contra-lateral, and bilateral ear effects have been reported in per­sons with brainstem lesions (Calearo & Antonelli, 1968; Lynn & Gilroy, 1977).

According to Lynn and Gilroy (1977), the diffuse nature of many brainstem lesions affects the correlation between locus of the tumor and the ear affected on the low-pass filtered speech test. The level of the lesion in the central nervous system and magnitude of the lesion also affect the correlation between locus of the tumor and the ear affected (Rintelmann, 1985).

The low-pass filtered speech test in the Willeford battery has a cutoff frequency of 500 Hz with a rejection rate of 18 dB/octave and uses Michigan consonant-nucleus-consonant (CNC) words. The presentation level is 50 dB SL relative to the PTA or SRT.
Because of the limited normative data available for the Willeford (CNC) and Lynn and Gilroy (NU-6) low-pass filtered speech test, lack of standardized procedure and tape, and the variety of cutoff frequencies available on commercial recordings of the filtered speech tests, each clinic should establish its own normative data.
B.

C. TIME COMPRESSED
Early investigations of time-compressed speech employed fast playback of the tape recorder. A drawback of the fast playback procedure was the concomitant shifts in spectrum of the recorded signal (Beasley & Freeman, 1977). The chop—splice procedure was then developed by Garvey (1953a,b), which corrected for the problem of frequency shifts associated with temporal alteration of the signal by manually cutting portions from the recording and then splicing back the remaining sections of the tape. This proce­dure, however, was tedious and awkward. Fairbanks, Everitt, and Jerger (1954) developed a procedure for time compression which avoided the problems associated with the playback and chop—splice procedure. Fairbanks et al. (1954) used an electromechanical time compressor/ expander which electro mechanically deleted samples and electromechanically spliced the remaining samples yielding a recording that is some percentage compressed or ex­panded relative to the original recording. The Springer In­formation Rate Changer was an electromechanical device similar to the Fairbanks device except that the discard inter­val had a limited range of variability. The Lee (1972) Vari-speech device, a modification of the Fairbanks instrument, contains a small tape recorder and minicomputer and is the one most widely used currently for time-compressed speech. A drawback of both the Fairbanks and Lee devices was that the sampling was random so samples discarded could be within as well as between linguistic sections. More sophisticated devices based on digital computers and speech synthesizers which allow for selectivity in removal of spe­cific sections are available but these devices are extremely expensive and are not employed clinically.

Calearo and Lazzaroni (1957) found that contralateral-ear effects were present in temporal-lobe tumor cases for time-compressed speech. Nevertheless, the contralateral-ear effects for time-compressed speech were not as pro­nounced as those for frequency-distorted speech. Calearo and Lazzaroni (1957) also reported that decreased supra-threshold speech-recognition scores for time-compressed speech were present for diffuse disorders of the central auditory nervous system and for brainstem pathology. Sim­ilar findings were obtained by deQuiros (1964).

Calearo and Antonelli (1968) reported that 14 of 23 cases with brainstem pathology had reduced scores on time-compressed speech tests, generally on one ear.

Kurdziel, Noffsinger, and Olsen (1976) obtained scores for 0, 40, and 60% time compression of NU-6 monosylla­bles for 15 patients with diffuse central nervous system lesions (e.g., cerebral vascular accident cases) and 16 pa­tients with discrete anterior temporal-lobe lesions. Contra-laterai-ear effects were present in the diffuse-lesion group but nor the discrete temporal-lobe lesion group. The contra-lateral-ear effects were enhanced at the highest compression rate (60%). Thus, the time-compressed speech is sensitive to diffuse hemispheric but not discrete anterior temporal-lobe lesions.

Beck and Mueller (1983) observed only slight contralateral-ear effects in patients with discrete brain in­jury resulting from head injury.
Studies of the effects of time compression on speech rec­ognition in normal-hearing elderly listeners enabled the in­vestigation of the effects of diffuse central nervous system dysfunction. Sticht and Gray (1969) investigated the effects of time compression of CID W-22 word lists on the su-prathreshold speech-recognition scores of young and old normal-hearing adults. Their results showed that older adults had more errors with time-compressed speech than younger adults at all degrees of time compression investi­gated.

Konkle, Beasley, and Bess (1977), who also used time-compressed W-22 word lists, obtained similar find­ings. The results indicate that time-compressed speech may help differentiate between peripheral and central auditory dysfunction in the elderly.

Beasley and his colleagues (1972a,b) have collected ex­tensive normative data for several rates of time compression and several sensation levels of presentation of NU-6 word lists using the Zemlin modification of the Fairbanks et al. (1954) procedure. Normative data were also obtained by deChicchis, Orchik, and Tecca (1981) for the Auditec uncompressed recordings of the NU-6 and W-22 word lists compressed using the Varispeech II time compressor/ expander. The scores reported by deChicchis et al. (1981) for the NU-6. word lists were significantly poorer than those reported by the Beasley et al. (1972a,b) norms. Beattie (1986) obtained normative performance-intensity func­tions for the Auditec CID W-22 recordings with 30 and 60% time compression.

A 30 and 60% time-compressed version of the NU-6 Form A tape was made commercially available by Auditec of St. Louis in 1978. Compression was obtained using the Varispeech II device. Grimes, Mueller, and Williams (1984)collected data on normal-hearing and sensorineural hearing-impaired persons for the four randomizations of the Auditec 60% time-compressed NU-6 Form A lists pre­sented at 32 dB SL relative to the SRT. Inter list equivalency was present only for lists I and IV in the sensorineural group and for lists 1, 11and IV in the normal-hearing group. In the normal-hearing group, the score reflecting rwo standard deviations below the mean of 73.2% was approximately 58% (across lists I, II, and IV). Grimes et al. suggested that interaural differences in time-compressed scores of at least 20% can be considered significant in normal-hearing listen­ers. Grimes et al. (1984) also reported the presence of sig­nificant learning effects which were larger in the normal-hearing group. Nevertheless, significant learning effects were not manifested between the first and second presen­tation. Grimes et al. (1984) concluded that the learning effect was clinically insignificant because not more than two lists are generally presented to a patient.
Beasley et al. (1972b) also reported significant list differ­ences in their recording of the NU-6 test. Riensche, Konkle, and Beasley (1976), who investigated interlist equivalency for the Beasley recording of the NU-6 Form A time-compressed lists while controlling for the order effect, re­ported a smaller interlist differences than reported by Bea­sley et al. (1972b) for Form B of the same recording.

Implication:

Comments

Popular posts from this blog

PERFORMANCE INTENSITY PHONETICALLY BALANCE (PIPB)

ALTERNATE BINAURAL LOUDNESS BALANCE TEST(ABLB)