Introduction 2005 1997 2001 1992 1990 1992 2003 2004 1992 2005 2003 2005 2006 2004 2004 2005 2005 2007 1982 2004 2005 2006 2000 2007 2006 2003 1997 Methods Subjects Twenty right-handed adult healthy subjects, native speakers of English, 8 women and 12 men, participated in the study. Subjects ranged in age from 23 to 28. All subjects gave informed consent to participate in the study. Experimental protocol 2005 In the “control” blocks the same physical deviants and standards as in the “oddball” blocks (deviant-counterparts and standard-counterparts, respectively), occurred quasi-randomly, while deviants were constrained by the same spacing rule mentioned above. However, each of the “control” blocks contained eight different equiprobable stimuli, including the deviant and standard counterparts. Thus, in the control runs the seven stimuli beside the deviant-counterpart served as “filler” or contextual stimuli which were added to the sequence so that the deviant-counterpart will appear at the same probability as any other stimulus in the sequence. Each of the stimuli in the “control” block (a total of eight different stimuli) repeated 42 times and appeared with equal probability which was identical to that of the deviants in the “oddball” block (12%). There were three blocks of each type (“oddball”/“control”). Each block was repeated twice. In total, 12 blocks of approximately 6 min each were randomly presented for each subject within a session. Three non-words, /de:/ “deh”; /te:/ “teh”; and /teI/ “tay”, and one word, /deI/ “day”, were recorded from a male native English speaker in a sound-proof chamber. These stimuli comprised the following standard-deviant pairs that resulted in the three “oddball” runs: (1) deh (standard)–day (deviant); (2) day–deh and (3) teh–tay. Thus, “deh” and “day” swapped their roles as standard and deviant in the second “oddball” block, whereas the third block controlled for the acoustic change associated with the transition from a monophthong (/e:/) to a diphthong (/eI/) occurring in the first standard-deviant pair. Additional four stimuli, /pe:/ “peh”; /peI/ “pay”; /be:/ “beh”; and /beI/ “bay”, together with the previous four mentioned above were embedded within three control runs, each containing a deviant-counterpart , either “day”, “deh” or “tay” appearing with the same probability and obeying the same spacing rule as in the oddball blocks. 1 http://www.praat.org Table 1 Pitch and frequencies of the main speech stimuli (in Hz) Stimulus F 0 F 1 F 1 Mean F 2 F 2 Mean F 3 F 3 Mean 1 101 554 634 603 1,714 1,667 1,702 2,548 2,617 2,636 2 99 539 640 624 1,726 1,593 1,658 2,576 2,583 2,636 3 101 539 630 614 1,730 1,600 1,666 2,585 2,590 2,635 1 104 477 406 411 2,076 2,230 2,142 2,693 2,706 2,703 2 102 492 412 419 1,983 2,137 2,089 2,645 2,669 2,668 3 103 497 462 451 2,018 2,063 2,051 2,684 2,703 2,685 1 101 581 527 537 1,858 1,911 1,942 2,581 2,495 2,641 2 102 610 566 600 1,787 1,887 1,906 2,671 2,635 2,670 3 103 600 589 622 1,790 1,839 1,915 2,660 2,622 2,681 1 103 636 623 709 1,709 1,662 1,747 2,587 2,625 2,695 2 101 783 626 707 1,750 1,650 1,735 2,584 2,525 2,639 3 101 857 615 747 1,805 1,608 1,748 2,563 2,531 2,655 F 0 F 1 F 2 F 3 F 1 F 2 F 3 The reason for using three tokens for each consonant-vowel stimulus was to control as much as possible for acoustic factors, other than those inherent in the structure of the stimulus, which could confound the elicitation of a deviant response. Using three different exemplars for each consonant-vowel (CV) stimulus diminished the likelihood of a contingency developing between a specific deviant-standard pair because of an uncontrolled acoustic facet associated with either the deviant or the standard. 1 F 1 F 2 1 2 2004 2 http://www.pstnet.com Table 2 “Oddball” and “control” sequences for the “deh” (standard)-“day” (deviant) pair Oddball deh1(A) day1(B) Control deh3(C) day2(D) A–D indicate the stimuli contrasted to evaluate: the deviance effect (B vs. A), the cognitive effect (B vs. D) and the sensory effect (A vs. C). The numbers attached to the stimuli indicate different exemplars of the same CV stimulus. Each contrast was computed across all exemplars of a specific CV. In the “oddball” sequence “day1” serves as a deviant and “deh1” functions as a standard. In the “control” condition “deh3” is a standard control counterpart and “day2” serves as a deviant control counterpart 1987 1990 1999 1997 2005 Behavioral assessment of discrimination between stimuli Thirteen subjects (out of the 20 participating in the study) rated the stimuli in a behavioral session performed outside of the magnet. The behavioral assessment was carried out in a different session. Stimuli were presented simultaneously with recorded MR scanner noise. The stimuli and the scanner noise were both presented in 62 dB SPL which was a convenient hearing level for both the stimuli and the noise presented together. Subjects were presented with “triplets” comprised of stimuli containing /e:/ and /eI/. In each block, 15 triplets were presented separated by a 2 s interval of silence allowing the subject to respond. The stimuli were the same as those presented in the fMRI experiment. There were several patterns of triplet presentation in each block, as follows. There were three triplets in which “day” was presented first followed by two “deh”-s; three triplets in which “day” was presented last and preceded by two “deh”-s; three triplets where “deh” was presented first followed by two “day”-s and three triplets were “deh” was last and preceded by two “day”-s. In addition, there were three triplets in which “day”, “deh” and “tay” each appeared in the middle position, respectively, while the neighboring stimuli (in the first and third positions) where either the /eI/ or /e:/ counterparts (i.e., “deh” “day” “deh”; “day” “deh” “day” and “teh “tay” “teh”). Overall, four blocks of 15 triplets each were presented to the subjects. The order of triplets was randomized within each block. The three different exemplars representing each CV stimulus were balanced across the four blocks. The subjects were required to indicate the outlier in each triplet by pressing the key (either “1”,”2” or “3” on the keyboard) that corresponded to the position of the outlier in the triplet. The outliers were assigned mainly to the extreme positions (1 or 3) in the triplet to simulate more closely the “oddball” design in which the deviant is surrounded by repetitive standards. The trials with the outlier appearing in the middle position were introduced to minimize the probability of guessing the identity of the third stimulus in the triplet after hearing the first two which were non-identical. Thus, subjects could guess the identity of the outlier with more confidence only after hearing two identical stimuli in a row. Data acquisition parameters Data was collected on a 3T Siemens Trio scanner. Each study began with two localizers: a three-plane localizer and a multiple-slice sagittal localizer. These were followed by the acquisition of twenty-five 6 mm T1-weighted axial slices (TR = 300 ms, TE = 2.47 ms, flip angle = 60°, FOV = 220 mm, 256 acquisition matrix). For each subject, 12 functional imaging scans were collected with slices in the same locations as the anatomical T1-weighted data. Functional images were recorded using a gradient-echo EPI sequence (TR = 1,550 ms, TE = 30 ms, flip angle = 80°, FOV = 220 mm, 64 acquisition matrix). Each functional run involved the acquisition of 245 volumes with twenty-five 6-mm axial slices. Images were converted to analyze format and the first ten volumes of each functional series were removed to account for the approach to steady-state magnetization, leaving 235 volumes for analysis. Image analysis Preprocessing x y z 2007 http://www.bioimagesuite.org/ 2006 2004 z x y 1998 Statistical analyses We used a two-stage random-effects model to analyze the data. In the first stage, statistical maps were calculated in the comparison of interest for each subject as described above. In the second stage, across subject analyses were conducted and the distribution of the individual subjects’ statistics were tested for significance. Within subject analyses Deviant maps Standard maps Deviant control maps Standard control Word map Non-word map Across subject analyses Composite maps. Paired t tests. t Deviant map Deviant control map t Standard map Standard control map 1995 P t P P Regions of interest analyses. t P n ROI analysis for the Word/Non-word maps. Word map Non-word map Word map Non-Word n Results Behavioral results t P 3 F P P Table 3 Reaction time (ms) in the behavioral triplet test Stimulus Position 1 2 3 day 787.58 ± 132.07 907.01 ± 240.71 638.55 ± 215.98 tay 858.59 ± 180.23 961.06 ± 187.15 650.98 ± 212.58 deh 864.03 ± 164.90 847.17 ± 158.27 628.33 ± 137.63 Reaction time is given for each of the three positions in the triplet and for each stimulus that served as a deviant in the “oddball” blocks. ±Standard deviation This confirms our prediction that reaction time will be shorter in case of an outlier presented in the third position following two identical stimuli. However, the fact that the main effect of Stimulus as well as the interaction between Stimulus and Position did not reach significance confirms that the stimuli were equally discriminable in the context of the “oddball” paradigms used in this experiment. Composite maps Deviance effect 1 1 Fig. 1 Deviant maps 1998 Upper panels left middle right Bottom panels two left panels two right panels z y x  4 4 Table 4 Deviant maps 1 Stimulus 3 t Talairach coordinates (mm) Side Identified region BA x y z day 10,803 4.72 56 −22 4 RT STG 22/42 day 11,349 4.73 −59 −21 6 LT STG 22/42 deh 3,562 4.91 55 −22 4 RT STG 21/22 deh 1,554 4.4 −62 −23 4 LT STG 22 tay 9,107 4.68 56 −18 2 RT STG 21/22 tay 6,034 4.97 −57 −18 3 LT STG 21/22 1988 t P STG RT LT BA 2005 Deviant control maps 2 2 5 Fig. 2 Deviant control maps Upper panels left middle right Bottom panels two left panels two right panels 1  Table 5 Deviant control maps 2 Stimulus 3 t Talairach coordinates (mm) Side Identified region BA x y z day 21,594 4.93 53 −16 15 RT PoCG/PrCG STG 40/43/6 day 27,919 4.93 −52 −18 16 LT PoCG/PrCG 40/43/6 day 4,600 −4.39 −24 48 30 LT SFrG 9 day 2,860 −4.2 −9 −80 −14 LT LT OCCP 18/19 deh 4,390 4.68 53 −21 9 RT STG 42 deh 5,021 4.4 −51 −21 6 LT STG 42 tay 3,401 4.6 −49 −8 38 LT PrCG 6 PoCG PrCG SFrG STG RT LT OCCP BA 4 2 5 2 5 t Cognitive effect Deviant maps Deviant control maps P 6 3 2005 Table 6 3 Stimulus 3 t Talairach coordinates (mm) Side Identified region BA x y z day 4,179 −4.37 41 −16 16 LT a 40/43 day 3,802 −4.5 −50 −20 19 RT a 40/43 LS INS a 4 Fig. 3 Deviant map Deviant control map 1998 z  Refractoriness effect t 4 7 4 2005 Fig. 4 Standard maps Standard control maps left middle 1998 z  Table 7 4 Stimulus 3 t Talairach coordinates (mm) Side Identified region BA x y z day 2,777 −4.27 59 −18 7 RT a 22/42 day 4,473 −4.39 −59 −20 4 LT a 22/42 deh 772 4.08 27 43 30 RT a 8/9 deh 2,510 4.29 −33 43 30 LT a 8/9 deh 3,189 4.31 3 14 42 Medial ACC 32 deh 2,726 −4.41 57 −26 3 RT STG 22 deh 3,030 −4.39 −57 −18 3 LT STG 22 tay 2,048 −4.35 59 −22 6 RT STG 22/42 tay 2,395 −4.53 −53 −17 4 LT STG 22 STG S/MFrG ACC a 4 4 7 3 6 2005 4 7 ROI analysis Discussion t t 3 t 4 t 4 7 t P t 5 5 Fig. 5 Anter_Aud Post_Aud Sup/Mid_Frontal  F P F P F P F P F P P 6 P within Fig. 6 Deviant maps Deviant-control maps 5  P between ROI analysis for the Word and Non-word maps Word map Non-word map 7 8 F P F P 8 P Fig. 7 Word and Non-Word maps Word map Non-word map Left panel Word-map Right panel Non-word map 1998 z  Table 8 Word Non-word maps 7 8 Stimulus 3 t Talairach coordinates (mm) Side Identified region BA x y z day (Word-map) 12,579 4.63 55 −16 14 RT PoCG/PrCG STG 40/43/6 day (Word-map) 14,260 4.65 −53 −18 13 LT STG 42 tay (Non-Word-map) 4,846 4.92 −52 −8 34 LT PrCG 4 PoCG PrCG STG 4 Fig. 8 Word Non-word 5  The distribution of percent signal change across subjects 9 Table 9 The distribution of percent signal change across subjects Brain region Side Map type n Mean ± SD % signal change LS/INS RT 3 17(−) −0.09 ± 0.07 LS/INS LT 3 18(−) −0.09 ± 0.10 STG RT 4 18(−) −0.15 ± 0.17 STG LT 4 17(−) −0.16 ± 0.14 S/MFrG RT 4 13(+) 0.06 ± 0.12 S/MFrG LT 4 13(+) 0.07 ± 0.10 PrCG LT 2 19(+) 0.13 ± 0.13 OCCP LT 2 18(−) −0.24 ± 0.20 t t LS INS STG S/MFrG PrCG OCCP RT LT No. of subj n Discussion Main findings 4 7 3 6 2005 2005 2004 2006 Support for the existence of “what” and “where” auditory streams 1996 2001 2002 2003 2002 Deviant 1 Deviant control 2 2000 2000 2003 2004 2003 Support for the existence of “what” and “where” auditory streams 2004 2006 2000 2004 2004 2006 The effect of speech perception Deviant maps 1 2 2003 2006 2 1994 2000 2 2 2006 5 2 Deviant control maps 2 5 2 5 4 7 5 2003 1993 2000 2003 2004 2004 2 2007 4 7 2005 2006 2006 2005 4 A possible effect of the lexical status in the “oddball condition 4 Deviant maps 2 4 1997 2003 1992 1996 Summary 2005 2007 2007 2005 2004 Conclusion 2004 2005 2004 2005 2006