Language comprehension in language-learning impaired children improved with acoustically modified speech

Science

Volumn 271, Issue 5245, 1996, Pages 81-84

  Tallal, Paula ^a   Miller, Steve L ^a   Bedi, Gail ^a   Byma, Gary ^a   Wang, Xiaoqin ^b   Nagarajan, Srikantan S ^b   Schreiner, Christoph ^b   Jenkins, William M ^b   Merzenich, Michael M ^b  

^a RUTGERS UNIVERSITY   (United States)

^b UNIVERSITY OF CALIFORNIA   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ACOUSTICS; ARTICLE; AUDITORY DISCRIMINATION; CHILD; CLINICAL ARTICLE; COMPREHENSION; COMPUTER PROGRAM; CONTROLLED STUDY; HUMAN; LANGUAGE DISABILITY; LEARNING DISORDER; PRESCHOOL CHILD; PRIORITY JOURNAL; SCHOOL CHILD; SPEECH; SPEECH DISCRIMINATION; TRAINING;

EID: 0030023111     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.271.5245.81     Document Type: Article

References (46)

1. P. K. Kuhl, K A Williams, F. Lacerda, K. N. Stevens, B. Lindblom, Science 255, 606(1992); P. D. Eimas, J L. Miller, P. S Jusczyk, in Categorical Perception, S Harnad, Ed (Cambridge Univ. Press, New York, 1987), pp 161-195, M. Steinschneider, J C. Arezzo, H. G Vaughan Jr., Brain Res. 519, 158 (1990)
2. P. K. Kuhl, K A Williams, F. Lacerda, K. N. Stevens, B. Lindblom, Science 255, 606(1992); P. D. Eimas, J L. Miller, P. S Jusczyk, in Categorical Perception, S Harnad, Ed (Cambridge Univ. Press, New York, 1987), pp 161-195, M. Steinschneider, J C. Arezzo, H. G Vaughan Jr., Brain Res. 519, 158 (1990)
3. P. K. Kuhl, K A Williams, F. Lacerda, K. N. Stevens, B. Lindblom, Science 255, 606(1992); P. D. Eimas, J L. Miller, P. S Jusczyk, in Categorical Perception, S Harnad, Ed (Cambridge Univ. Press, New York, 1987), pp 161-195, M. Steinschneider, J C. Arezzo, H. G Vaughan Jr., Brain Res. 519, 158 (1990)
4. J H. Beitchman, R. Nair, M. Clegg, P. G. Patel, J. Speech Hear. Disord 51, 98 (1986); P A. Silva, Dev. Med Child Neurol. 22, 768 (1980), J. Stevenson and N. Richman, ibid. 18, 431 (1976); D M Williams, J. D. Derbyshire, D A Vaghy, J. Otolaryngol. 7 (suppl.), 5 (1980)
5. J H. Beitchman, R. Nair, M. Clegg, P. G. Patel, J. Speech Hear. Disord 51, 98 (1986); P A. Silva, Dev. Med Child Neurol. 22, 768 (1980), J. Stevenson and N. Richman, ibid. 18, 431 (1976); D M Williams, J. D. Derbyshire, D A Vaghy, J. Otolaryngol. 7 (suppl.), 5 (1980)
6. J H. Beitchman, R. Nair, M. Clegg, P. G. Patel, J. Speech Hear. Disord 51, 98 (1986); P A. Silva, Dev. Med Child Neurol. 22, 768 (1980), J. Stevenson and N. Richman, ibid. 18, 431 (1976); D M Williams, J. D. Derbyshire, D A Vaghy, J. Otolaryngol. 7 (suppl.), 5 (1980)
7. J H. Beitchman, R. Nair, M. Clegg, P. G. Patel, J. Speech Hear. Disord 51, 98 (1986); P A. Silva, Dev. Med Child Neurol. 22, 768 (1980), J. Stevenson and N. Richman, ibid. 18, 431 (1976); D M Williams, J. D. Derbyshire, D A Vaghy, J. Otolaryngol. 7 (suppl.), 5 (1980)
8. Proceedings of the National Conference on Learning Disabilities, Bethesda, MD, 12 and 13 January 1987 (York, Parkton, MD, 1988)
9. R E. Stark et al , Ann Dyslexia 34, 49 (1984), D. V. M. Bishop and C. Adams, J Child Psychol Psychiatry 31, 1027 (1990), H W. Catts, J. Speech Hear Res 36, 948 (1993), I. Y. Liberman, D. Shankweiler, R. W. Fischer, B. Carter, J Exp. Child Psychol 18, 201 (1974); H. S. Scarborough, Child Dev. 61, 1728 (1990); J. K. Torgesen, R K. Wagner, C. A. Rashotte, J Learning Disabil 5, 276 (1994).
10. R E. Stark et al , Ann Dyslexia 34, 49 (1984), D. V. M. Bishop and C. Adams, J Child Psychol Psychiatry 31, 1027 (1990), H W. Catts, J. Speech Hear Res 36, 948 (1993), I. Y. Liberman, D. Shankweiler, R. W. Fischer, B. Carter, J Exp. Child Psychol 18, 201 (1974); H. S. Scarborough, Child Dev. 61, 1728 (1990); J. K. Torgesen, R K. Wagner, C. A. Rashotte, J Learning Disabil 5, 276 (1994).
11. R E. Stark et al , Ann Dyslexia 34, 49 (1984), D. V. M. Bishop and C. Adams, J Child Psychol Psychiatry 31, 1027 (1990), H W. Catts, J. Speech Hear Res 36, 948 (1993), I. Y. Liberman, D. Shankweiler, R. W. Fischer, B. Carter, J Exp. Child Psychol 18, 201 (1974); H. S. Scarborough, Child Dev. 61, 1728 (1990); J. K. Torgesen, R K. Wagner, C. A. Rashotte, J Learning Disabil 5, 276 (1994).
12. R E. Stark et al , Ann Dyslexia 34, 49 (1984), D. V. M. Bishop and C. Adams, J Child Psychol Psychiatry 31, 1027 (1990), H W. Catts, J. Speech Hear Res 36, 948 (1993), I. Y. Liberman, D. Shankweiler, R. W. Fischer, B. Carter, J Exp. Child Psychol 18, 201 (1974); H. S. Scarborough, Child Dev. 61, 1728 (1990); J. K. Torgesen, R K. Wagner, C. A. Rashotte, J Learning Disabil 5, 276 (1994).
13. R E. Stark et al , Ann Dyslexia 34, 49 (1984), D. V. M. Bishop and C. Adams, J Child Psychol Psychiatry 31, 1027 (1990), H W. Catts, J. Speech Hear Res 36, 948 (1993), I. Y. Liberman, D. Shankweiler, R. W. Fischer, B. Carter, J Exp. Child Psychol 18, 201 (1974); H. S. Scarborough, Child Dev. 61, 1728 (1990); J. K. Torgesen, R K. Wagner, C. A. Rashotte, J Learning Disabil 5, 276 (1994).
14. R E. Stark et al , Ann Dyslexia 34, 49 (1984), D. V. M. Bishop and C. Adams, J Child Psychol Psychiatry 31, 1027 (1990), H W. Catts, J. Speech Hear Res 36, 948 (1993), I. Y. Liberman, D. Shankweiler, R. W. Fischer, B. Carter, J Exp. Child Psychol 18, 201 (1974); H. S. Scarborough, Child Dev. 61, 1728 (1990); J. K. Torgesen, R K. Wagner, C. A. Rashotte, J Learning Disabil 5, 276 (1994).
15. M Rissman, S Curtiss, P. Tallal, J Speech-Lang. Pathol. Aud. 14, 49 (1990).
16. P Tallal and M Percy, Nature 241, 468, (1973).
17. _, Neuropsychologia 12, 83 (1974); P. Tallal, Brain Lang. 9, 182 (1980); _ and R. E Stark, Ann Dyslexia 32, 163 (1982) , _, E. D. Mellits, Brain Lang. 25, 314 (1985); R. Stark and P. Tallal, Language, Speech and Reading Disorders in Children. Neuropsychological Studies (College-Hill Press, Boston, 1988).
18. _, Neuropsychologia 12, 83 (1974); P. Tallal, Brain Lang. 9, 182 (1980); _ and R. E Stark, Ann Dyslexia 32, 163 (1982) , _, E. D. Mellits, Brain Lang. 25, 314 (1985); R. Stark and P. Tallal, Language, Speech and Reading Disorders in Children. Neuropsychological Studies (College-Hill Press, Boston, 1988).
19. _, Neuropsychologia 12, 83 (1974); P. Tallal, Brain Lang. 9, 182 (1980); _ and R. E Stark, Ann Dyslexia 32, 163 (1982) , _, E. D. Mellits, Brain Lang. 25, 314 (1985); R. Stark and P. Tallal, Language, Speech and Reading Disorders in Children. Neuropsychological Studies (College-Hill Press, Boston, 1988).
20. _, Neuropsychologia 12, 83 (1974); P. Tallal, Brain Lang. 9, 182 (1980); _ and R. E Stark, Ann Dyslexia 32, 163 (1982) , _, E. D. Mellits, Brain Lang. 25, 314 (1985); R. Stark and P. Tallal, Language, Speech and Reading Disorders in Children. Neuropsychological Studies (College-Hill Press, Boston, 1988).
21. _, Neuropsychologia 12, 83 (1974); P. Tallal, Brain Lang. 9, 182 (1980); _ and R. E Stark, Ann Dyslexia 32, 163 (1982) , _, E. D. Mellits, Brain Lang. 25, 314 (1985); R. Stark and P. Tallal, Language, Speech and Reading Disorders in Children. Neuropsychological Studies (College-Hill Press, Boston, 1988).
22. P. Tallal and M Piercy, Neuropsychologia 13, 69 (1975).
23. M. M. Merzenich et al., Science 271, 77 (1996).
24. Speech modification was achieved by a two-stage processing algorithm. In the first stage, the rate of the speech signal was prolonged by 50%, while preserving its spectral content and natural quality. This time scale modification was implemented with a digital signal processing algorithm [M. R Portnoff, IEEE Trans. Acoust. Speech Signal Process 29 (no. 3), 374 (1981)]. This algonthm involved computation of the short-time Founer transform (STFT) of the speech signal with the fast-Fourier transform (FFT), linear interpolation, and phase-modification of the STFT to the new time scale, followed by additive synthesis with the inverse-Fourier transform. In the second stage of processing, the fast transition elements were differentially amplified by as much as 20 dB. The fast transition elements of speech were defined as the 3- to 30-Hz components of the speech envelope within rate-changed narrow-band channels This differential "emphasis" was also implemented with a digital signal processing algorithm. The modification involved band pass filtering of the speech signal into critical-band channels, computation of the envelope within each channel, band-pass filtenng of the speech envelope, modification of the narrow-band signals to carry the new band-pass envelope followed by additive synthesis of the modified speech signal from the narrow-band channels. The above-mentioned algonthm was implemented initially by using a filter-bank summation algorithm and then improved on with the overlap-add procedure and the FFT [T. Langhans and I. I. W. Strube, Proc. of IEEE-International Conference on Acoustics and Speech Signal Processing 1982, 156 (1982)].
25. Speech modification was achieved by a two-stage processing algorithm. In the first stage, the rate of the speech signal was prolonged by 50%, while preserving its spectral content and natural quality. This time scale modification was implemented with a digital signal processing algorithm [M. R Portnoff, IEEE Trans. Acoust. Speech Signal Process 29 (no. 3), 374 (1981)]. This algonthm involved computation of the short-time Founer transform (STFT) of the speech signal with the fast-Fourier transform (FFT), linear interpolation, and phase-modification of the STFT to the new time scale, followed by additive synthesis with the inverse-Fourier transform. In the second stage of processing, the fast transition elements were differentially amplified by as much as 20 dB. The fast transition elements of speech were defined as the 3- to 30-Hz components of the speech envelope within rate-changed narrow-band channels This differential "emphasis" was also implemented with a digital signal processing algorithm. The modification involved band pass filtering of the speech signal into critical-band channels, computation of the envelope within each channel, band-pass filtenng of the speech envelope, modification of the narrow-band signals to carry the new band-pass envelope followed by additive synthesis of the modified speech signal from the narrow-band channels. The above-mentioned algonthm was implemented initially by using a filter-bank summation algorithm and then improved on with the overlap-add procedure and the FFT [T. Langhans and I. I. W. Strube, Proc. of IEEE-International Conference on Acoustics and Speech Signal Processing 1982, 156 (1982)].
26. M. M. Merzenich and W. M. Jenkins, in Maturational Windows and Adult Cortical Plasticity, B. Julesz and I Kovécs, Eds. (Addison-Wesley, Reading, MA, 1995), pp. 247-264
27. The following five tests were used as clinical benchmark measures dunng pretraimng (week 1) and posttraining (week 6) These clinical speech and language tests were recorded with natural, unmodified speech and presented over headphones, (i) F. DiSimoni, The Token Test for Children (Teaching Resources Corporation, Boston, MA, 1978). The Token Test assesses the ability to follow auditory commands of increasing length and grammatical complexity, (ii) R. Goldman, M Fnstoe, R. W. Woodcock, Goldman-Fristoe-Woodcock Diagnostic Auditory Discrimination Test (American Guidance Service, Circle Pines, MN, 1974). The GFW test assesses speech-sound discrimination within words, (iii) S Curtiss and J Yamada, Curtiss and Yamada Comprehensive Language Evaluation-Receptive (CYCLE-R), unpublished. The CYCLE-R thoroughly examines comprehension of specific components of grammar (morphology and syntax), (iv) P. Tallal and S Miller, Computerized Version of the Tallal Repetition Test, unpublished (1994). The Computerized Repetition Test is a modification of the Repetition Test [P. Tallal, in Non-Speech Language and Communication, R. Schiefelbusch, Ed. (University Park Press, Baltimore, MD, 1980), pp. 449-467)]. In the Repetition Test, subjects are operantly trained to press one panel on a response box after hearing stimulus 1 and a different panel for stimulus 2 Two stimuli are then presented sequentially in various combinations (that is, 1-1, 2-1, 1-2, and 2-2) with an interstimulus interval (|S|) interposed between the two tones The subject is required to reproduce the sequence by pressing the panels in the correct order. The Computerized Repetition Test determines the threshold |S| at which sequences of two pure-tone stimuli of 150-, 75-, 40-, or 17-ms duration are perceived and reproduced with 75% accuracy. The |S|s vary from 500 to 0 ms. (v) R. Goldman and M. Fnstoe, Goldman-Fristoe Test of Articulation (American Guidance Service, Circle Pines, MN, 1986). The Soundsin-Words subtest was used to assess accuracy in speech articulation. Speech was elicited by having the child label a picture that depicted a common object or activity
28. The following five tests were used as clinical benchmark measures dunng pretraimng (week 1) and posttraining (week 6) These clinical speech and language tests were recorded with natural, unmodified speech and presented over headphones, (i) F. DiSimoni, The Token Test for Children (Teaching Resources Corporation, Boston, MA, 1978). The Token Test assesses the ability to follow auditory commands of increasing length and grammatical complexity, (ii) R. Goldman, M Fnstoe, R. W. Woodcock, Goldman-Fristoe-Woodcock Diagnostic Auditory Discrimination Test (American Guidance Service, Circle Pines, MN, 1974). The GFW test assesses speech-sound discrimination within words, (iii) S Curtiss and J Yamada, Curtiss and Yamada Comprehensive Language Evaluation-Receptive (CYCLE-R), unpublished. The CYCLE-R thoroughly examines comprehension of specific components of grammar (morphology and syntax), (iv) P. Tallal and S Miller, Computerized Version of the Tallal Repetition Test, unpublished (1994). The Computerized Repetition Test is a modification of the Repetition Test [P. Tallal, in Non-Speech Language and Communication, R. Schiefelbusch, Ed. (University Park Press, Baltimore, MD, 1980), pp. 449-467)]. In the Repetition Test, subjects are operantly trained to press one panel on a response box after hearing stimulus 1 and a different panel for stimulus 2 Two stimuli are then presented sequentially in various combinations (that is, 1-1, 2-1, 1-2, and 2-2) with an interstimulus interval (|S|) interposed between the two tones The subject is required to reproduce the sequence by pressing the panels in the correct order. The Computerized Repetition Test determines the threshold |S| at which sequences of two pure-tone stimuli of 150-, 75-, 40-, or 17-ms duration are perceived and reproduced with 75% accuracy. The |S|s vary from 500 to 0 ms. (v) R. Goldman and M. Fnstoe, Goldman-Fristoe Test of Articulation (American Guidance Service, Circle Pines, MN, 1986). The Soundsin-Words subtest was used to assess accuracy in speech articulation. Speech was elicited by having the child label a picture that depicted a common object or activity
29. The following five tests were used as clinical benchmark measures dunng pretraimng (week 1) and posttraining (week 6) These clinical speech and language tests were recorded with natural, unmodified speech and presented over headphones, (i) F. DiSimoni, The Token Test for Children (Teaching Resources Corporation, Boston, MA, 1978). The Token Test assesses the ability to follow auditory commands of increasing length and grammatical complexity, (ii) R. Goldman, M Fnstoe, R. W. Woodcock, Goldman-Fristoe-Woodcock Diagnostic Auditory Discrimination Test (American Guidance Service, Circle Pines, MN, 1974). The GFW test assesses speech-sound discrimination within words, (iii) S Curtiss and J Yamada, Curtiss and Yamada Comprehensive Language Evaluation-Receptive (CYCLE-R), unpublished. The CYCLE-R thoroughly examines comprehension of specific components of grammar (morphology and syntax), (iv) P. Tallal and S Miller, Computerized Version of the Tallal Repetition Test, unpublished (1994). The Computerized Repetition Test is a modification of the Repetition Test [P. Tallal, in Non-Speech Language and Communication, R. Schiefelbusch, Ed. (University Park Press, Baltimore, MD, 1980), pp. 449-467)]. In the Repetition Test, subjects are operantly trained to press one panel on a response box after hearing stimulus 1 and a different panel for stimulus 2 Two stimuli are then presented sequentially in various combinations (that is, 1-1, 2-1, 1-2, and 2-2) with an interstimulus interval (|S|) interposed between the two tones The subject is required to reproduce the sequence by pressing the panels in the correct order. The Computerized Repetition Test determines the threshold |S| at which sequences of two pure-tone stimuli of 150-, 75-, 40-, or 17-ms duration are perceived and reproduced with 75% accuracy. The |S|s vary from 500 to 0 ms. (v) R. Goldman and M. Fnstoe, Goldman-Fristoe Test of Articulation (American Guidance Service, Circle Pines, MN, 1986). The Soundsin-Words subtest was used to assess accuracy in speech articulation. Speech was elicited by having the child label a picture that depicted a common object or activity
30. The following five tests were used as clinical benchmark measures dunng pretraimng (week 1) and posttraining (week 6) These clinical speech and language tests were recorded with natural, unmodified speech and presented over headphones, (i) F. DiSimoni, The Token Test for Children (Teaching Resources Corporation, Boston, MA, 1978). The Token Test assesses the ability to follow auditory commands of increasing length and grammatical complexity, (ii) R. Goldman, M Fnstoe, R. W. Woodcock, Goldman-Fristoe-Woodcock Diagnostic Auditory Discrimination Test (American Guidance Service, Circle Pines, MN, 1974). The GFW test assesses speech-sound discrimination within words, (iii) S Curtiss and J Yamada, Curtiss and Yamada Comprehensive Language Evaluation-Receptive (CYCLE-R), unpublished. The CYCLE-R thoroughly examines comprehension of specific components of grammar (morphology and syntax), (iv) P. Tallal and S Miller, Computerized Version of the Tallal Repetition Test, unpublished (1994). The Computerized Repetition Test is a modification of the Repetition Test [P. Tallal, in Non-Speech Language and Communication, R. Schiefelbusch, Ed. (University Park Press, Baltimore, MD, 1980), pp. 449-467)]. In the Repetition Test, subjects are operantly trained to press one panel on a response box after hearing stimulus 1 and a different panel for stimulus 2 Two stimuli are then presented sequentially in various combinations (that is, 1-1, 2-1, 1-2, and 2-2) with an interstimulus interval (|S|) interposed between the two tones The subject is required to reproduce the sequence by pressing the panels in the correct order. The Computerized Repetition Test determines the threshold |S| at which sequences of two pure-tone stimuli of 150-, 75-, 40-, or 17-ms duration are perceived and reproduced with 75% accuracy. The |S|s vary from 500 to 0 ms. (v) R. Goldman and M. Fnstoe, Goldman-Fristoe Test of Articulation (American Guidance Service, Circle Pines, MN, 1986). The Soundsin-Words subtest was used to assess accuracy in speech articulation. Speech was elicited by having the child label a picture that depicted a common object or activity
31. The following five tests were used as clinical benchmark measures dunng pretraimng (week 1) and posttraining (week 6) These clinical speech and language tests were recorded with natural, unmodified speech and presented over headphones, (i) F. DiSimoni, The Token Test for Children (Teaching Resources Corporation, Boston, MA, 1978). The Token Test assesses the ability to follow auditory commands of increasing length and grammatical complexity, (ii) R. Goldman, M Fnstoe, R. W. Woodcock, Goldman-Fristoe-Woodcock Diagnostic Auditory Discrimination Test (American Guidance Service, Circle Pines, MN, 1974). The GFW test assesses speech-sound discrimination within words, (iii) S Curtiss and J Yamada, Curtiss and Yamada Comprehensive Language Evaluation-Receptive (CYCLE-R), unpublished. The CYCLE-R thoroughly examines comprehension of specific components of grammar (morphology and syntax), (iv) P. Tallal and S Miller, Computerized Version of the Tallal Repetition Test, unpublished (1994). The Computerized Repetition Test is a modification of the Repetition Test [P. Tallal, in Non-Speech Language and Communication, R. Schiefelbusch, Ed. (University Park Press, Baltimore, MD, 1980), pp. 449-467)]. In the Repetition Test, subjects are operantly trained to press one panel on a response box after hearing stimulus 1 and a different panel for stimulus 2 Two stimuli are then presented sequentially in various combinations (that is, 1-1, 2-1, 1-2, and 2-2) with an interstimulus interval (|S|) interposed between the two tones The subject is required to reproduce the sequence by pressing the panels in the correct order. The Computerized Repetition Test determines the threshold |S| at which sequences of two pure-tone stimuli of 150-, 75-, 40-, or 17-ms duration are perceived and reproduced with 75% accuracy. The |S|s vary from 500 to 0 ms. (v) R. Goldman and M. Fnstoe, Goldman-Fristoe Test of Articulation (American Guidance Service, Circle Pines, MN, 1986). The Soundsin-Words subtest was used to assess accuracy in speech articulation. Speech was elicited by having the child label a picture that depicted a common object or activity
32. The following five tests were used as clinical benchmark measures dunng pretraimng (week 1) and posttraining (week 6) These clinical speech and language tests were recorded with natural, unmodified speech and presented over headphones, (i) F. DiSimoni, The Token Test for Children (Teaching Resources Corporation, Boston, MA, 1978). The Token Test assesses the ability to follow auditory commands of increasing length and grammatical complexity, (ii) R. Goldman, M Fnstoe, R. W. Woodcock, Goldman-Fristoe-Woodcock Diagnostic Auditory Discrimination Test (American Guidance Service, Circle Pines, MN, 1974). The GFW test assesses speech-sound discrimination within words, (iii) S Curtiss and J Yamada, Curtiss and Yamada Comprehensive Language Evaluation-Receptive (CYCLE-R), unpublished. The CYCLE-R thoroughly examines comprehension of specific components of grammar (morphology and syntax), (iv) P. Tallal and S Miller, Computerized Version of the Tallal Repetition Test, unpublished (1994). The Computerized Repetition Test is a modification of the Repetition Test [P. Tallal, in Non-Speech Language and Communication, R. Schiefelbusch, Ed. (University Park Press, Baltimore, MD, 1980), pp. 449-467)]. In the Repetition Test, subjects are operantly trained to press one panel on a response box after hearing stimulus 1 and a different panel for stimulus 2 Two stimuli are then presented sequentially in various combinations (that is, 1-1, 2-1, 1-2, and 2-2) with an interstimulus interval (|S|) interposed between the two tones The subject is required to reproduce the sequence by pressing the panels in the correct order. The Computerized Repetition Test determines the threshold |S| at which sequences of two pure-tone stimuli of 150-, 75-, 40-, or 17-ms duration are perceived and reproduced with 75% accuracy. The |S|s vary from 500 to 0 ms. (v) R. Goldman and M. Fnstoe, Goldman-Fristoe Test of Articulation (American Guidance Service, Circle Pines, MN, 1986). The Soundsin-Words subtest was used to assess accuracy in speech articulation. Speech was elicited by having the child label a picture that depicted a common object or activity
33. The speech and language exercises were developed as games to maintain attention and motivation over the course of the study Tape recorded syllables, words, phrases, and sentences that had been acoustically modified with the speech algorithm developed for this study were presented to the child over headphones or free field. The games included acting out commands in a Simon Says format with props; pointing to pictures or colored blocks in response to commands; repeating verbatim syllables, nonsense words, real words, or sentences; and pointing to pictures corresponding to spoken words. Throughout training, commands of increasing length and grammatical complexity were used in these games Careful attention was given in the design of the listening exercises to ensure that foils developed for each item would focus the attention of the child on the salient aspects of speech discrimination or receptive grammar being trained. In the listening games, regardless of the accuracy of the child's response, immediate nonverbal feedback was given after each response ("thumbs up" or "thumbs down"), followed by a repetition of the item with the correct response indicated by the clinician, so the child could have a second chance to process correctly Each child won points for cooperation throughout the training, which were tallied daily and exchanged for pnzes at the end of each week.
34. Changes from study 1 to study 2 included (i) increasing the duration of the laboratory sessions from 3 to 3.5 hours per day, (ii) providing homework solely in the form of recorded children's stories on tape [either acoustically modified (group A) or with natural unmodified speech (group B)] instead of computer games, (iii) increasing the number of computer game formats from two to four, and (iv) modifying the ratio of clinicians to children in each training session from one-to-one to usually one-to-one, but on occasion one-to-two. The children in study 1 and group A in study 2 received computer games that adaptively trained temporal processing and phoneme perception, whereas the children in group B study 2 received the same schedule of computer game training and reinforcement, but with games that did not contain temporally or phonetically adaptive stimuli.
35. Subjects were assigned to the two groups to minimize the differences between subjects on measures of performance IQ (PIQ) [Wechsler Intelligence Seals for Children-III (The Psychological Corporation, New York, 1991)] reported as mean (SEM) [PIQ, group A = 96 1 (2 6), group B = 96.6 (3.3)], and receptive language performance (Token Test Age scores) reported as mean (SEM) [group A = 5.4 (0.4), group B = 6 1 (0 7)].
36. Previous studies [P. Tallal and M. Piercy, Neuropsychologia 11, 389 (1973)] have shown that the total signal duration of auditory stimulus patterns, as indexed by the relation between the duration and interval among stimulus elements, is critical for demonstrating the temporal processing deficits of LLI children. In the present investigation, temporal threshold values were calculated as the sum of the minimal tone durations (150-, 75-, 40-, or 17-ms tone pairs) and the average |S| based on an adaptive staircase (two-up and one-down) procedure to which subjects were able to reproduce pairs of tone sequences by pressing a response panel. A performance level of 75% or greater accuracy was required at a particular stimulus duration before a threshold would be calculated The average pretraining thresholds by the LLI children were 491 ms in study 1 and 287 ms in study 2 (9) Normally developing children of a comparable age have been shown to require |S|s of less than 20 ms on this test (5).
37. P. Tallal, R. E. Stark, D Mellits, Neuropsychologia 23, 527 (1985b).
38. S. Curtiss, W. Katz, P. Tallal, Am Speech Hear. Assoc. 35, 373 (1992); L. B Leonard, Appl. Psycholinguist. 10, 179 (1989).
39. S. Curtiss, W. Katz, P. Tallal, Am Speech Hear. Assoc. 35, 373 (1992); L. B Leonard, Appl. Psycholinguist. 10, 179 (1989).
40. Six weeks after training was completed in study 1, six of the seven children were retested with the same battery of benchmark speech and language measures to determine the extent to which the significant gains made between pre- and posttraining were maintained, without further exposure to acoustically modified speech. The results showed that the significant improvements over pretraining baseline scores were maintaned.
41. A. A. Benasich and P. Tallal, Infant Behav. Dev., in press; in Temporal Information Processing in the Nervous System: Special Reference to Dyslexia and Dysphasia, P. Tallal, A. M Galaburda, R. R Llinás, C von Euler, Eds (New York Academy of Sciences, New York, 1993), vol 682, pp 312-314; J. L Henderson and S. E. Trehub, paper presented at the 61 st Biennial Meeting of the Society for Research in Child Development, Indianapolis, IN, 1 April 1995.
42. A. A. Benasich and P. Tallal, Infant Behav. Dev., in press; in Temporal Information Processing in the Nervous System: Special Reference to Dyslexia and Dysphasia, P. Tallal, A. M Galaburda, R. R Llinás, C von Euler, Eds (New York Academy of Sciences, New York, 1993), vol 682, pp 312-314; J. L Henderson and S. E. Trehub, paper presented at the 61 st Biennial Meeting of the Society for Research in Child Development, Indianapolis, IN, 1 April 1995.
43. A. A. Benasich and P. Tallal, Infant Behav. Dev., in press; in Temporal Information Processing in the Nervous System: Special Reference to Dyslexia and Dysphasia, P. Tallal, A. M Galaburda, R. R Llinás, C von Euler, Eds (New York Academy of Sciences, New York, 1993), vol 682, pp 312-314; J. L Henderson and S. E. Trehub, paper presented at the 61 st Biennial Meeting of the Society for Research in Child Development, Indianapolis, IN, 1 April 1995.
44. S. Gordon-Salant and P. J. Fitzgibbons, J. Speech Hear Res 36, 1276 (1993); P. Tallal and F. New-combe, Brain and Lang. 5, 13 (1978).
45. S. Gordon-Salant and P. J. Fitzgibbons, J. Speech Hear Res 36, 1276 (1993); P. Tallal and F. New-combe, Brain and Lang. 5, 13 (1978).
46. Informed consent was obtained from the parent or parents of each child after the potential risks and benefits of the studies were explained. We thank the therapists who referred subjects as well as the parents and children who participated. We thank A. Rubenstein, B. Glazewski, J. Flax, C. Roesler, K Masters, J. Reitzel, T. Delaney, and P. Johnston for assistance in subject selection, stimulus preparation, and clinical testing and T Realpe, I. Shell, C. Kapelyan, A Katsnelson, L Brzustowicz, C. Brown, A. Khoury, and S Shapack for assistance in the experimental training. Valuable comments on the manuscript by I. Creese are appreciated. We thank the Charles A. Dana Foundation for supporting the research. For more information, see http://www.ld.ucsf.edu