Deviance detection in sound frequency in simple and complex sounds in urethane-anesthetized rats

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY-NC-ND 4.0 https://creativecommons.org/licenses/by-nc-nd/4.0/ Deviance detection in sound frequency in simple and complex sounds in urethaneanesthetized rats © 2019 Elsevier B.V. All rights reserved. Accepted version (Final draft) Yang, Tiantian; Hämäläinen, Jarmo; Lohvansuu, Kaisa; Lipponen, Arto; Penttonen, Markku; Astikainen, Piia Yang, T., Hämäläinen, J., Lohvansuu, K., Lipponen, A., Penttonen, M., & Astikainen, P. (2021). Deviance detection in sound frequency in simple and complex sounds in urethane-anesthetized rats. Hearing Research, 399, Article 107814. https://doi.org/10.1016/j.heares.2019.107814 2021 Journal Pre-proof Deviance detection in sound frequency in simple and complex sounds in urethaneanesthetized rats Tiantian Yang, Jarmo Hämäläinen, Kaisa Lohvansuu, Arto Lipponen, Markku Penttonen, Piia Astikainen PII: S0378-5955(19)30270-9 DOI: https://doi.org/10.1016/j.heares.2019.107814 Reference: HEARES 107814 To appear in: Hearing Research Received Date: 11 June 2019 Revised Date: 4 October 2019 Accepted Date: 9 October 2019 Please cite this article as: Yang, T., Hämäläinen, J., Lohvansuu, K., Lipponen, A., Penttonen, M., Astikainen, P., Deviance detection in sound frequency in simple and complex sounds in urethaneanesthetized rats, Hearing Research, https://doi.org/10.1016/j.heares.2019.107814. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Elsevier B.V. All rights reserved. 1 1 Deviance detection in sound frequency in simple and complex sounds in 2 urethane-anesthetized rats 3 4 Tiantian Yang, Jarmo Hämäläinen, Kaisa Lohvansuu, Arto Lipponen, Markku Penttonen, 5 Piia Astikainen * 6 7 Affiliation of all authors: Department of Psychology, University of Jyvaskyla, Jyväskylä, 8 Finland; 9 10 Address: Department of Psychology, P.O. Box 35, 40014 University of Jyväskylä, Finland 11 12 * Corresponding author: 13 Piia Astikainen 14 Phone: +358408053480 15 facsimile: +358142601021 16 e-mail: [email protected] 17 18 Tiantian Yang: tiantian.t.ya[email protected]yu.fi 19 Jarmo Hämäläinen: [email protected] 20 Kaisa Lohvansuu: [email protected] 21 Arto Lipponen: [email protected] 22 Markku Penttonen: [email protected] 23 24 2 Abstract 25 26 Mismatch negativity (MMN), which is an electrophysiological response demonstrated in 27 humans and animals, reflects memory-based deviance detection in a series of sounds. 28 However, only a few studies on rodents have used control conditions that were sufficient in 29 eliminating confounding factors that could also explain differential responses to deviant 30 sounds. Furthermore, it is unclear if change detection occurs similarly for sinusoidal and 31 complex sounds. In this study, we investigated frequency change detection in 32 urethane-anesthetized rats by recording local-field potentials from the dura above the 33 auditory cortex. We studied change detection in sinusoidal and complex sounds in a series of 34 experiments, controlling for sound frequency, probability, and pattern in a series of sounds. 35 For sinusoidal sounds, the MMN controlled for frequency, adaptation, and pattern, was 36 elicited at approximately 200 ms onset latency. For complex sounds, the MMN controlled for 37 frequency and adaptation, was elicited at 60 ms onset latency. Sound frequency affected the 38 differential responses. MMN amplitude was larger for the sinusoidal sounds than for the 39 complex sounds. These findings indicate the importance of controlling for sound frequency 40 and stimulus probabilities, which have not been fully controlled for in most previous animal 41 and human studies. Future studies should confirm the preference for sinusoidal sounds over 42 complex sounds in rats. 43 44 45 Keywords: auditory cortex, change detection, local-field potentials, mismatch negativity, rat 46 3 47 4 Abbreviations 48 CSAS: Cascade ascending control condition 49 MMN: Mismatch negativity 50 MST: Many-standards control condition 51 NMDA: N-methyl-D-aspartate 52 5 53 1. Introduction 54 Detecting changes in sensory environment is important for adaptive behavior. Brain 55 responses that reflect change detection are studied using so-called oddball stimulus condition 56 in which a rare “deviant” stimulus is interspersed with a frequent “standard” stimulus. 57 Recordings of the brain’s electrical responses to oddball stimuli elicit the mismatch negativity 58 (MMN) component (Näätänen et al., 1978). The MMN in the auditory domain has been well 59 documented in both humans and animals (Harms et al., 2016; Näätänen et al., 2010), and it 60 has the potential for clinical applications (Näätänen, 2000; 2003). 61 62 MMN is a response to a rare stimulus that violates the regularity formed by a repetitive 63 standard stimulus (Näätänen et al., 2005). According to the memory-comparison hypothesis 64 the brain compares an incoming (deviant) stimulus with the memory trace formed by a 65 standard stimulus. When these two are found to differ, MMN is elicited (Näätänen et al., 66 2005). Another possible explanation for differential responses to deviant stimuli (in 67 comparison to standard stimuli) is the different levels of adaptation in the neural populations 68 responding to frequent and rare sounds because the neurons that respond to standard sound 69 features are more frequently activated than are the neurons that respond to deviant sound 70 features (May & Tiitinen, 2010). 71 72 Different control conditions have been developed to study whether mere adaptation can 73 explain differential responses elicited in the oddball condition (Harms et al., 2016). Currently, 74 6 the most frequently used control condition is the many-standards condition, which is also 75 termed the equal probability condition (Ruhnau et al., 2012; Schröger & Wolff, 1996). It 76 contains several other stimuli in addition to the deviant and standard stimuli in the oddball 77 condition. In the many-standards condition, stimuli are presented otherwise in a random order, 78 but without consecutive repetitions of the same sound. Also, the probability of each sound is 79 the same as the probability of the oddball-deviant sound. Therefore, the oddball-deviant 80 stimulus and control stimulus differ only by the context in which they are presented: 81 regularity formed by a standard stimulus in the oddball condition vs. continuously changing 82 stimuli without regularity in the many-standards condition. Importantly, because the 83 oddball-deviant stimulus and its control sound have the same probability of occurrence, the 84 responses they elicit include the same level of adaptation. In humans, in the many-standards 85 control condition, it has been demonstrated that the differential response found in the oddball 86 condition cannot be explained by adaptation only (Jacobsen & Schröger, 2001; Jacobsen & 87 Schröger, 2003; Jacobsen et al., 2003; Lohvansuu et al., 2013; Maess et al., 2007) because 88 oddball-deviant sound elicits larger responses than the physically identical control sound. 89 90 Another appropriate but less used control condition is the cascade condition, which comprises 91 a fixed pattern of stimuli (Ruhnau et al., 2012). When frequency change detection is 92 investigated, the stimuli are ordered according to ascending and/or descending frequency, and 93 the repeated pattern allows the upcoming stimulus to be predictable. This feature is 94 comparable to the oddball condition in which repeated standard stimuli are predictable. 95 96 7 Rodent MMN studies using frequency changes in stimuli have applied either the 97 many-standards (e.g. Astikainen et al., 2011; Farley et al., 2010; Fishman & Steinschneider, 98 2012; Nakamura et al., 2011; Polterovich et al., 2018; Shiramatsu et al., 2013) or cascade 99 (Harms et al., 2014; Parras et al., 2017) control conditions. A potential confounder in some 100 studies is that smaller frequency differences were applied between the sounds in the control 101 condition than between the stimuli in the oddball condition (Astikainen et al., 2011; Fishman 102 & Steinschneider, 2012; Jung et al., 2013; Kurkela et al., 2018; Nakamura et al., 2011). Thus, 103 the results may have been confounded because of different levels of across-frequency 104 adaptation, a mechanism that reduces responses to the consecutively presented sounds of 105 nearby frequencies (Taaseh et al., 2011), between the oddball and control condition sounds. 106 Moreover, the control conditions in previous rodent studies were different from those used in 107 the original human studies. In previous human studies (Jacobsen & Schröger, 2001; Maess et 108 al., 2007; Ruhnau et al., 2012; Wiens et al., 2019), the control sounds were at one end of the 109 frequency range. In previous rodent studies, the control sounds were positioned in the middle 110 of other sounds (Astikainen et al., 2011; Farley et al., 2010; Fishman & Steinschneider, 2012; 111 Harms et al., 2014; Nakamura et al., 2011; Parras et al., 2017; Taaseh et al., 2011). It is 112 possible that the across-frequency adaptation on the responses differed from sounds that were 113 assigned to the end of the frequency range compared to those in the middle position. 114 115 The aim of the present study is to investigate change detection of sound frequency in 116 anesthetized rats. The MMN responses were recorded in the dura above the auditory cortex 117 for two sound frequencies: 4000 Hz and 4600 Hz. Both sound frequencies were assigned as 118 14 The responses to the deviant sounds and those to the same frequency tone that served as the 247 standard in the oddball sequence were compared using a paired, two-tailed t-test. The 248 sample-point-by-sample-point pairs were compared from the stimulus onset to the end of the 249 sweeps. To counteract type 1 errors due to multiple comparisons, an alpha level of 0.05 was 250 required in at least 20 consecutive data points (10 ms) (Ruusuvirta et al., 2015). If such robust 251 differential responses were found in the oddball condition, the responses to the deviant tone 252 and standard tone were further compared to the same tones in the many-standards and 253 cascade-ascending conditions. A 10-ms period of significant difference was again set as the 254 minimum for a robust amplitude difference between the deviant and control sound responses. 255 256 PLEASE INSERT FIGURE 1 APPROXIMATELY HERE 257 258 3. Results 259 All the sounds elicited the typical waveform reported in studies using urethane-anesthetized 260 rats (e.g., Astikainen et al., 2011), where a positive polarity deflection peaked at 261 approximately 35 ms, and a negative-polarity deflection peaked at approximately 80 ms after 262 the stimulus onset. In the present study, the MMN was calculated between the deviant and 263 standard responses elicited by the same frequency sounds. However, the responses to the 264 standard and deviant responses obtained separately from each oddball condition comparing 265 different frequency standard and deviant responses are reported in the supplementary material 266 (Supplement 1). 267 268 15 269 3.1. Responses to sinusoidal sounds 270 The 4000 Hz deviant sound shifted the responses toward negative polarity in comparison to 271 the 4000 Hz standard sound at 204.5–253.5 ms post-stimulus latency (Figure 2, Table 1). In 272 this time range, the responses to the 4000 Hz deviant sound were more negative than to the 273 corresponding control sound in the cascade condition. The same result was found in the 274 comparison between the 4000 Hz deviant and the corresponding control sound in the 275 many-standards condition, but the difference began to appear at 215.5 ms after stimulus 276 onset. 277 278 Regarding the sinusoidal 4600 Hz tones, the responses to the deviant and the standard sounds 279 differed at 56.5–86.5 ms latency (Figure 2, Table 1). The deviant sound shifted the responses 280 toward negative polarity compared with the responses to the standard sounds. There were no 281 differences between the responses to the deviant and control sounds during the same time 282 period. The response to the 4600 Hz standard tone was smaller (i.e., more positive) than the 283 response to the same tone in the cascade control condition at 67.5–84.5 ms latency. 284 285 PLEASE INSERT FIGURE 2 APPROXIMATELY HERE 286 287 3.2. Responses to complex sounds 288 16 The response to the 4000 Hz deviant sound was larger toward negative polarity than the 289 response to the same frequency standard tone at 64.5–84.5 ms latency (Figure 3, Table 1). 290 The responses to the deviant sound were more negative in amplitude than the responses to the 291 control sound in the many-standards condition at 64.5–84.0 ms latency, but no difference was 292 found between the responses to the deviant and the control sounds in the cascade condition. 293 294 Regarding the analysis of repetition suppression, there was no difference between the 295 responses to the 4000 Hz standard sound and the corresponding sound in the control 296 conditions. 297 298 No differential response was found to the complex 4600 Hz sound. 299 300 301 PLEASE INSERT FIGURE 3 APPROXIMATELY HERE 302 303 3.3. Comparison of the differential responses to sinusoidal and complex sounds 304 The differential responses between the deviant and standard sounds of the same frequency 305 were compared using point-by-point t-tests (Figure 4). The differential responses to the 4000 306 Hz sounds were larger (i.e., greater negative polarity) for the sinusoidal sounds than for the 307 complex sounds at 209–238 ms after the stimulus onset. The differential responses to the 308 4600 Hz sounds were also larger (i.e., greater negative polarity) in the sinusoidal than in the 309 complex sounds at 55.5–85 ms latency. 310 17 311 PLEASE INSERT FIGURE 4 APPROXIMATELY HERE 312 Table 1. Latencies of the significant differences (in ms). 313 Differential response in Oddball Genuine MMN (MST) Genuine MMN (CSAS) Repetition suppression (MST) Repetition suppression (CSAS) Sinusoidal 4000 Hz 204.5 – 253.5 215.5 – 253.5 204.5 – 253.5 n.s. n.s. Sinusoidal 4600 Hz 56.5 – 86.5 n.s. n.s. n.s. 67.5 – 84.5 Complex 4000 Hz 64.5 – 84.5 64.5 – 84 n.s n.s. n.s. Complex 4600 Hz n.s. n.a. n.a. n.a. n.a. Note. Differential response (deviant-standard) was calculated as a difference between the responses to the same 314 frequency sounds obtained in the two oddball conditions. Genuine MMN = Genuine mismatch negativity 315 latency (i.e., when the deviant stimulus responses were larger than the responses to the control stimulus). In the 316 repetition suppression analyses, standard responses were compared with the same frequency sounds of the 317 many-standards (MST) and cascade-ascending (CSAS) control conditions: n.s. = non-significant; n.a. = not 318 applicable (not calculated because there was no differential response in the oddball condition). 319 320 18 4. Discussion 321 The present study was designed to investigate change detection of sound frequency as 322 reflected by a genuine MMN defined here as a differential response to a rare deviant sound 323 that reflects the detection of regularity violations. The adaptation effect, which can affect the 324 responses because of lower probability and thus smaller neural adaptation related to rare than 325 frequent sounds (May & Tiitinen, 2010), was controlled by applying two control conditions: 326 the many-standards condition and the cascade ascending condition. 327 328 In this study, the MMN was calculated as the difference between the standard and deviant 329 responses elicited by the same frequency sounds (e.g., deviant 4000 Hz - standard 4000 Hz). 330 Sound frequency affected the differential responses, however. When the differential response 331 was calculated for the 4600 Hz sounds, an early latency differential response (starting at 56.5 332 ms post-stimulus) was found. However, it was not a genuine MMN because the response to 333 the deviant stimulus in the oddball condition did not differ from the response to the same 334 frequency sound in the control conditions. Therefore, the differential response in the oddball 335 condition cannot be considered a genuine MMN. Instead, the differential response could be 336 explained by different amounts of neural adaptation to standard and deviant stimuli due to 337 their different repetition rates in the stimulus series. Furthermore, repetition suppression was 338 found in this condition: the responses to standards were more adapted than the responses to 339 the cascade control sounds. No difference was found between the responses to the standard 340 sounds and those to the control sounds in the many-standards condition. However, a different 341 result was found for the differential response calculated between the 4000 Hz sinusoidal 342 19 sounds. The differential response between deviant and standard responses was elicited 343 starting at 204.5 ms post-stimulus latency, and it reflected the detection of regularity 344 violations (i.e., a genuine MMN) because the amplitude was higher for the responses to the 345 deviant sounds than for the responses to the control sounds. 346 347 The pattern of the results differed regarding the complex sounds, which elicited an early 348 (starting at 64.5 ms post-stimulus) genuine MMN to the 4000 Hz sounds. Similar to the 349 sinusoidal sounds, the complex sounds elicited no differential response to the 4600 Hz sounds. 350 Regarding the 4000 Hz complex sounds, the oddball-deviant responses and control sound 351 responses differed only when the control sound was obtained in the many-standards control 352 condition but not when the control sound response was obtained in the cascade condition. 353 These results indicated that a genuine MMN was not found when the predictive pattern of the 354 sound sequences was controlled. 355 356 In summary, a genuine MMN was found for both sinusoidal and complex sounds of 4000 Hz, 357 but not of 4600 Hz frequency. Therefore, the stimulus frequency had a significant effect on 358 the MMN elicitation, which was observed in the recording electrode on the dura above the 359 auditory cortex. This finding aligns with previous findings in studies on 360 urethane-anesthetized rats, which showed differential responses to deviant sounds of some 361 but not all sound frequencies (Harms et al., 2014; Nakamura et al., 2011; Ruusuvirta et al., 362 2015). In Ruusuvirta et al. (2015), eight different deviant sound frequencies were interspersed 363 with a standard sound. The statistical model also showed that the sound frequency, not the 364 20 stimulus probability or context, explained the differential responses to the oddball-deviant 365 sounds (Ruusuvirta et al., 2015). In the present study, the sound frequency was controlled by 366 applying the flip-flop condition, and the MMN was calculated between the same frequency 367 standard and deviant sounds. The MMN was elicited to 4000 Hz sounds but not to the 4600 368 Hz sounds. This result could be explained by the position of the recording electrode, which 369 could have been more favorable for the mismatch response to the 4000 Hz sounds than to the 370 4600 Hz sounds. The reason for this could be that the frequency of the test sound guiding the 371 electrode position was 1661 Hz. A lower frequency test sound comparing to the sinusoidal 372 sound frequencies applied in the oddball condition was used because the aim was to make the 373 recording location suitable for both sinusoidal and complex sounds. The latter contained 374 lower frequencies than the ones applied in the sinusoidal sounds. However, because we did 375 not record responses from different auditory cortical areas simultaneously (e.g., Shiramatsu et 376 al., 2013), the ultimate reason for differential responses to 4000 Hz, but not to 4600 Hz, 377 deviant sounds remains unclear. In future research, the measurement of local-field potentials 378 with multiple electrodes over the auditory cortex should be combined with an analysis in 379 which MMN is calculated as the difference between the responses to the same frequency 380 standard and deviant sounds. 381 382 The results showed that the complex sounds elicited the controlled MMN in early (i.e., before 383 100 ms) latency and the sinusoidal sounds in later (i.e., after 200 ms) latency. The earlier 384 response (MMN before 100 ms latency) has been frequently reported in rats and mice as a 385 genuine MMN controlled for adaptation to speech sounds (Ahmed et al., 2011) and 386 21 sinusoidal sounds (Astikainen et al., 2011; Kurkela et al., 2018; Nakamura et al., 2011; Parras 387 et al., 2017; Polterovich et al., 2018). However, some previous results may have been 388 confounded by the sound frequency because the flip-flop condition was not applied (Ahmed 389 et al., 2011; Astikainen et al., 2011; Kurkela et al., 2018) or the same frequency standard and 390 deviant responses were not compared (Parras et al., 2017; Polterovich et al., 2018). Similar to 391 our study, Nakamura et al. (2011) compared the responses to the same frequency sounds 392 obtained in the two flip-flop conditions in rats. They found genuine MMN for high (3600 Hz) 393 frequency sounds but not for low (2500 Hz) frequency sounds. In contrast, our results showed 394 genuine MMN in lower but not in higher frequency sounds. This difference can be explained 395 by the electrodes’ position, which could have been more optimal for one frequency than the 396 other. 397 398 The later response at 200 ms latency, which was a genuine MMN, was elicited only by 4000 399 Hz sinusoidal sounds; that is, not by 4600 Hz sinusoidal sounds or complex sounds of either 400 frequency. Late MMN latencies have been observed less often in rodents, but in our previous 401 mice study, a genuine MMN with a large effect size was found until 255 ms latency in the 402 descending deviant frequency and 185 ms latency in the ascending deviant frequency 403 (Kurkela et al., 2018). In a previous study that investigated stimulus-specific adaptation (SSA) 404 in rats to oddball sounds, a late response at 200–400 ms latency was found in the excitatory 405 neurons (Chen et al., 2015). Importantly, this late response shared some of the defining 406 features of the MMN: it was larger for oddball sounds than for control sounds, and it was 407 reduced when the intracellular N-methyl-D-aspartate (NMDA) receptors were inhibited. Our 408 22 result showing a late MMN at over 200 ms latency thus extends the previous similar findings 409 of local-field potentials in mice and single cell responses in rats to local-field potentials in 410 rats. Further studies are required to elucidate the possible functional roles of early and late 411 latency MMN in rodents. 412 413 The direct comparison of the differential responses to sinusoidal and complex sounds (Fig. 4) 414 showed that the sinusoidal sounds elicited larger differential responses than the complex 415 sounds did before 100 ms latency (4600 Hz sounds) and after 200 ms latency (4000 Hz 416 sounds). This finding was surprising because previous studies on humans showed that 417 complex sounds elicited larger brain responses than sinusoidal sounds did (Tervaniemi, 418 Schröger et al., 2000; Tervaniemi, Ilvonen et al., 2000). The discrepancy between the current 419 rat study and earlier human studies is likely due to the methodological differences in these 420 studies. In the present study, the brain responses were measured directly from the dura of the 421 auditory cortex. Neural generators of complex sound responses are different from those of 422 sinusoidal sound responses (Alho et al., 1996; Novitski et al., 2004). Therefore, our recording 423 electrode could have been better placed to detect the responses to the sinusoidal sounds rather 424 than the complex sounds. Further studies using multiple recording sites in the auditory cortex 425 in rats are required to clarify this issue. Another reason for the contradictory results could be 426 that the human brain responds better to complex sounds than to sinusoidal sounds because it 427 is specialized and/or more exposed to speech sounds, but there is no such preference in the rat 428 brain. 429 430 23 The methods applied in the present study included controls that were more stringent than 431 those used in previous animal studies. However, genuine MMN was elicited by both 432 sinusoidal and complex sounds in the anesthetized rats. The pattern of results indicate the 433 importance of controlling for sound frequency, which may be particularly important when a 434 single intracranial recording electrode is used to study frequency change detection. Although 435 local-field potentials recorded in the dura capture activity in a large area, the tonotopic 436 organization of the auditory cortex can affect MMN responses, especially when it is 437 calculated as the difference between the deviant and standard responses of different 438 frequencies. Therefore, the most valid way to define the MMN is to calculate the difference 439 in the responses to the standard and deviant stimuli that are of the same frequency. However, 440 only a few human MMN studies have applied this calculation method (Jacobsen & Schröger, 441 2001; Maess et al., 2007). It is possible that human scalp-recorded event-related potentials are 442 not as sensitive as the intracranial local-field potentials to sound frequency because the signal 443 that is recorded over the scalp reflects activity of a large cortical area. However, similar 444 control conditions and analyses used to detect physically identical deviant and control sounds 445 would increase the comparability of rodent and human MMN studies. 446 447 In the present study, the pattern of the results for the MMN response to sinusoidal sounds was 448 the same regardless of whether the many-standards or the cascade-ascending condition was 449 applied. Repetition suppression in comparison to standard sounds was found only in 450 comparison to the 4600 Hz control sound in the cascade condition but not in the 451 many-standards condition. Regarding the complex sounds, the control responses obtained in 452 30 Näätänen, R. (2000). Mismatch negativity (MMN): Perspectives for application. 566 International Journal of Psychophysiology, 37(1), 3-10. 567 Näätänen, R. (2003). Mismatch negativity: Clinical research and possible applications. 568 International Journal of Psychophysiology, 48(2), 179-188. 569 Näätänen, R., Astikainen, P., Ruusuvirta, T., & Huotilainen, M. (2010). Automatic auditory 570 intelligence: An expression of the sensory–cognitive core of cognitive processes. Brain 571 Research Reviews, 64(1), 123-136. 572 Näätänen, R., Gaillard, A. W., & Mäntysalo, S. (1978). Early selective-attention effect on 573 evoked potential reinterpreted. Acta Psychologica, 42(4), 313-329. 574 Näätänen, R., Jacobsen, T., & Winkler, I. (2005). Memory based or afferent processes in ‐ 575 mismatch negativity (MMN): A review of the evidence. Psychophysiology, 42(1), 25-32. 576 Nakamura, T., Michie, P. T., Fulham, W. R., Todd, J., Budd, T. W., Schall, U., . . . Hodgson, 577 D. M. (2011). Epidural auditory event-related potentials in the rat to frequency and 578 duration deviants: Evidence of mismatch negativity? Frontiers in Psychology, 2, 367. 579 Novitski, N., Tervaniemi, M., Huotilainen, M., & Näätänen, R. (2004). Frequency 580 discrimination at different frequency levels as indexed by electrophysiological and 581 behavioral measures. Cognitive Brain Research, 20(1), 26-36. 582 31 Parras, G. G., Nieto-Diego, J., Carbajal, G. V., Valdés-Baizabal, C., Escera, C., & Malmierca, 583 M. S. (2017). Neurons along the auditory pathway exhibit a hierarchical organization of 584 prediction error. Nature Communications, 8(1), 2148. 585 Polterovich, A., Jankowski, M. M., & Nelken, I. (2018). Deviance sensitivity in the auditory 586 cortex of freely moving rats. PloS One, 13(6), e0197678. 587 Ruhnau, P., Herrmann, B., & Schröger, E. (2012). Finding the right control: The mismatch 588 negativity under investigation. Clinical Neurophysiology, 123(3), 507-512. 589 Ruusuvirta, T., Koivisto, K., Wikgren, J., & Astikainen, P. (2007). Processing of melodic 590 contours in urethane anaesthetized rats.‐ European Journal of Neuroscience, 26(3), 591 701-703. 592 Ruusuvirta, T., Lipponen, A., Pellinen, E., Penttonen, M., & Astikainen, P. (2015). Auditory 593 cortical and hippocampal local-field potentials to frequency deviant tones in 594 urethane-anesthetized rats: An unexpected role of the sound frequencies themselves. 595 International Journal of Psychophysiology, 96(3), 134-140. 596 Schröger, E., & Wolff, C. (1996). Mismatch response of the human brain to changes in sound 597 location. Neuroreport, 7(18), 3005-3008. 598 Shiramatsu, T. I., Kanzaki, R., & Takahashi, H. (2013). Cortical mapping of mismatch 599 negativity with deviance detection property in rat. PLoS One, 8(12), e82663. 600 32 Taaseh, N., Yaron, A., & Nelken, I. (2011). Stimulus-specific adaptation and deviance 601 detection in the rat auditory cortex. PLoS One, 6(8), e23369. 602 Tervaniemi, M., Ilvonen, T., Sinkkonen, J., Kujala, A., Alho, K., Huotilainen, M., & 603 Näätänen, R. (2000). Harmonic partials facilitate pitch discrimination in humans: 604 Electrophysiological and behavioral evidence. Neuroscience Letters, 279(1), 29-32. 605 Tervaniemi, M., Schröger, E., Saher, M., & Näätänen, R. (2000). Effects of spectral 606 complexity and sound duration on automatic complex-sound pitch processing in 607 humans–a mismatch negativity study. Neuroscience Letters, 290(1), 66-70. 608 Wiens, S., Szychowska, M., Eklund, R., & van Berlekom, E. (2019). Cascade and 609 no-repetition rules are comparable controls for the auditory frequency mismatch 610 negativity in oddball tasks. Psychophysiology, 56(1), e13280. doi:10.1111/psyp.13280 611 612 613 614 33 Figure 1. The illustration of the stimulus conditions. Here, sinusoidal conditions are used as 615 an example, but all the stimulus conditions were also applied to complex sounds. A) Oddball 616 descending condition: a 4000 Hz tone is a deviant stimulus, and a 4600 Hz tone is a standard 617 stimulus. B) Oddball ascending condition: the assignment of the standard and deviant 618 stimulus is reversed compared to the oddball descending condition. C) Many-standards 619 control condition: in consistently changing stimuli, the control stimuli were 4000 Hz and 620 4600 Hz tones. D) Cascade-ascending control condition: fixed pattern of stimuli, control 621 stimuli were 4000 Hz and 4600 Hz tones. In both control conditions, the probability of each 622 tone was 0.125 (as for the deviant in the oddball condition), and similar to the oddball 623 condition, there was a 15% frequency difference between the tones. 624 625 Figure 2. Responses to the sinusoidal sounds (above 4000 Hz and below 4600 Hz). The 626 shadings of the waveforms represent 95% confidence intervals (CI). The stimulus onset was 627 at time 0. The left column: standard (STD) and deviant (DEV) responses to sounds of the 628 same frequency (responses obtained in the flip-flop condition) and a differential response 629 (deviant–standard). The black horizontal bars in the bottom of the left column figures mark 630 the time window of the significant difference between the deviant and standard responses: 95% 631 CI is presented for the standard and deviant stimulus responses. The middle column: 632 responses to the deviant stimulus and the same frequency stimulus in the many-standards 633 (MST) and cascade-ascending control conditions (CSAS): 95% CI is presented for the MST 634 and CSAS responses. The green and yellow horizontal bars mark the time window of the 635 significant difference between the responses to the deviant sounds and each control sound. 636 34 The right column: responses to the oddball standard and the same frequency stimulus in the 637 many-standard and cascade control condition: 95% CI is presented for the MST and CSAS 638 responses. The yellow horizontal bars mark the time window of the significant difference 639 between the deviant and each control sound response. Oddball and control responses were 640 compared only when there was a significant difference between the responses to the standard 641 and deviant stimuli in the oddball condition. 642 643 Figure 3. Responses to the complex sounds (above 4000 Hz and below 4600 Hz). The 644 shadings of the waveforms represent 95% CI. The stimulus onset was at time 0. The left 645 column: standard and deviant responses to sounds at the same frequency (responses obtained 646 in the flip-flop condition) and a differential response (deviant–standard): 95% CI is presented 647 for the standard and deviant stimulus responses. The black horizontal bars at the bottom of 648 the left column figures mark the time window of significant difference between the deviant 649 and standard responses. The middle column: responses to deviant stimulus and to the same 650 frequency stimulus in the many-standards (MST) and cascade-ascending (CSAS) control 651 conditions: 95% CI is presented for the MST and CSAS responses. The green horizontal bar 652 marks the time window of the significant difference between deviant responses and each 653 control sound response. The right column: responses in the standard, many-standards, and 654 cascade control conditions to the same frequency sounds: 95% CI is presented for the MST 655 and CSAS responses. The oddball and control responses were compared only when there was 656 35 a significant difference between the responses to the standard and deviant stimuli in the 657 oddball condition. 658 659 Figure 4. Differential responses (deviant–standard) to sinusoidal and complex sounds. 660 Differential responses to 4000 Hz (left) and 4600 Hz (right) sounds with shading show 95% 661 CI. The stimulus onset was at time 0. The black horizontal bars show the latency of the 662 significant difference between the differential responses to the sinusoidal and complex 663 sounds. 664 665