Consecutive TMS-fMRI reveals remote effects of neural noise to the “occipital face area”

Author’s Accepted Manuscript Consecutive TMS-fMRI Reveals Remote Effects of Neural Noise to the “Occipital Face Area” Lily M. Solomon-Harris, Sara A. Rafique, Jennifer K.E. Steeves www.elsevier.com/locate/brainres

PII: DOI: Reference:

S0006-8993(16)30604-7 http://dx.doi.org/10.1016/j.brainres.2016.08.043 BRES45088

To appear in: Brain Research Received date: 8 December 2015 Revised date: 12 August 2016 Accepted date: 29 August 2016 Cite this article as: Lily M. Solomon-Harris, Sara A. Rafique and Jennifer K.E. Steeves, Consecutive TMS-fMRI Reveals Remote Effects of Neural Noise to the “Occipital Face Area”, Brain Research, http://dx.doi.org/10.1016/j.brainres.2016.08.043 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. 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.

1

Consecutive TMS-fMRI Reveals Remote Effects of Neural Noise to the “Occipital Face Area”

Lily M. Solomon-Harrisa,b, Sara A. Rafiquea,b, Jennifer K.E. Steevesa,b,* a

b

Centre for Vision Research and Department of Psychology, York University, Toronto, Canada

Centre for Vision Research and Department of Psychology, York University, 4700 Keele St.,

Toronto, Ontario, M3J 1P3, Canada *

Corresponding author: Jennifer Steeves 1032 Sherman Health Science Research Centre York

University 4700 Keele Street, Toronto, Ontario, M3J 1P3 Canada. Tel. 416-736-2100 Ext. 20452; Fax 416-736-5814. [email protected]

Abstract The human cortical system for face perception comprises a network of connected regions including the middle fusiform gyrus (“fusiform face area” or FFA), the inferior occipital gyrus (“occipital face area” or OFA), and the posterior superior temporal sulcus (pSTS). Here, we sought to investigate how transcranial magnetic stimulation (TMS) to the OFA affects activity within the face processing network. We used offline repetitive TMS to temporarily introduce neural noise in the right OFA in healthy subjects. We then immediately performed functional magnetic resonance imaging (fMRI) to measure changes in blood oxygenation level dependent (BOLD) signal across the face network using an fMR-adaptation paradigm. We hypothesized that TMS to the right OFA would induce abnormal face identity coding throughout the face processing network in regions to which it has direct or indirect connections. Indeed, BOLD

2 signal for face identity, but not non-face (butterfly) identity, decreased in the right OFA and FFA following TMS to the right OFA compared to both sham TMS and TMS to a control site, the nearby object-related lateral occipital area (LO). Further TMS to the right OFA decreased facerelated activation in the left FFA, without any effect in the left OFA. Our findings indicate that TMS to the right OFA selectively disrupts face coding at both the stimulation site and bilateral FFA. TMS to the right OFA also decreased BOLD signal for different identity stimuli in the right pSTS. Together with mounting evidence from patient studies, we demonstrate connectivity of the OFA within the face network and that its activity modulates face processing in bilateral FFA as well as the right pSTS. Moreover, this study shows that remote deep regions within the face network can be probed by stimulating structures closer to the cortical surface.

Keywords: connectivity; face perception; FFA; fusiform gyrus; inferior occipital gyrus; OFA

1. Introduction The perception of faces by humans relies on a specialized neural network, one that is distinct from other forms of visual processing (e.g. Kanwisher, 2000). The relationship between regions in this network continues to be described by researchers. The purpose of the present study is to probe the role of the inferior occipital gyrus or “occipital face area” (OFA; Gauthier et al., 2000) within the face perception network. We used transcranial magnetic stimulation (TMS) to introduce neural noise within the OFA and then consecutively using functional magnetic resonance imaging (fMRI) we measured changes in blood oxygenation level dependent (BOLD) signal at the stimulation site and other connected areas.

3 Hierarchical feedforward models posit that face processing follows a similar mechanism as other types of visual processing, where basic feature detection is followed by increasingly sophisticated and global analysis (e.g. Ungerleider and Mishkin, 1982). That is, the basic features of a face are first detected in posterior visual areas, with information passing from early visual cortex to the OFA, followed by integration of this information into a more global representation in more anterior brain regions, including the middle fusiform gyrus or “fusiform face area” (FFA; Kanwisher et al., 1997) for coding identity, and the posterior part of the superior temporal sulcus (pSTS; Puce et al., 1998) for analysis of viewpoint and expression (e.g. Haxby et al., 2000; Pitcher et al., 2011b). There is evidence, however, that this network is not purely feedforward or hierarchical in nature and may instead have several feedback loops. For example, the OFA codes more than just basic information since both the OFA and FFA are sensitive to the spatial relations of facial features (Rhodes et al., 2009) and both are involved in the coding of face identity (Xu and Biederman, 2010). Patient data also support a more complex interplay between these regions of the face network. Two prosopagnosia patients, DF and PS, who have an inability to visually recognize faces, have common lesions at the right OFA, but nonetheless show face-selective activation at the right FFA. This suggests the possibility of a more direct flow of information to the FFA from early visual cortex or perhaps subcortically, bi-passing the OFA (Johnson, 2005; Milner et al., 1991; Rossion et al., 2003; Steeves et al., 2006). Despite retaining this faceselective activation, the sensitivity to face identity is abnormal in the right FFA of both patients, evidence that an intact OFA is integral for higher-level identity coding in the FFA (Steeves et al., 2009).

4 Others suggest a model of face perception where the more posterior OFA, with its smaller receptive fields, engages in fine-grained feature analysis after the more anterior FFA, with its larger receptive fields, has holistically categorized a stimulus as a face (Rossion, 2008; Steeves et al., 2009). By providing this detailed analysis following initial categorization, the OFA thus greatly facilitates subsequent identity coding (Jiang et al., 2015; Jiang et al., 2011; Rossion et al., 2011; Rossion et al., 2012). Rossion’s (2008) model is consistent with evidence of extensive bi-directional cortical connections (Felleman and Van Essen, 1991), the reverse hierarchy theory of visual perception (Hochstein and Ahissar, 2002), and the ability of prosopagnosia patients to categorize faces with a specific deficit in higher-level face recognition processes (Steeves et al., 2006). In a previous study, we used TMS to introduce neural noise within the right OFA, and found categorization of intact compared to scrambled faces was unaffected, but face identification was significantly impaired (Solomon-Harris et al., 2013). These findings are consistent with the notion that the FFA categorizes faces prior to input from the OFA to enable to face identification. We demonstrated that this recognition impairment was specific to faces, and further, that stimulation of a nearby region, the right lateral occipital area (LO), did not impair face recognition. In the present study we further investigate the role of the OFA in the face-processing network and examine the effects of the introduction of neural noise within the OFA on BOLD signal in the FFA and other connected areas. We performed consecutive TMS-fMRI (e.g. Mullin and Steeves, 2013; Rafique et al., 2015) with offline repetitive low frequency TMS outside the scanner (targeting the OFA and the object-preferential LO, in separate sessions) followed by functional neuroimaging to measure BOLD signal across the face network. The consecutive

5 TMS-fMRI paradigm has three parts: (1) prestimulation fMRI to functionally localize the TMS targets and regions-of-interest (ROIs) for later fMRI analysis; (2) the application of TMS to the functionally defined targets (in separate sessions on different days); and (3) poststimulation fMRI to examine effects of TMS on BOLD signal. Immediately following TMS, we employed an fMR-adaptation (fMR-A) paradigm with images of faces and butterflies of different and repeated identities, similar to previous studies of acquired prosopagnosia (Figure 1; Dricot et al., 2008; Steeves et al., 2009). In fMR-A, BOLD signal of a given brain region is reduced following repeated presentations of a given stimulus property, but recovers from this adapted state by an increase in BOLD signal with changes in the given stimulus property. This effect is termed release from adaptation and indicates that the region codes that stimulus property (Grill-Spector and Malach, 2001). We predicted that TMS to the OFA would reduce face identity coding, reflected by a decreased adaptation response in both the OFA and FFA.

2. Results 2.1 Data Analysis R statistical computing software was used for all analyses (The R Project for Statistical Computing, www.r-project.org). After verifying that the assumptions for analysis of variance (ANOVA) were satisfied, a 3x2x2 [Stimulation Condition (TMS to OFA, TMS to LO, Sham) x Image Category (Face or Butterfly) x Image Type (Same or Different)] omnibus repeated measures ANOVA was performed on the BOLD signal (beta weights) for each ROI (Figures 2 and 3). Significant findings and a priori comparisons of interest (i.e., effects of TMS at the stimulation sites and the remote FFA) were followed up with t-tests for pairwise comparisons,

6 and p-values were adjusted with the false discovery rate (FDR; Benjamini & Hochberg, 1995) correction for multiple comparisons. Alpha was set at p<0.05 for significant differences and p<0.10 for trends. Effect sizes were calculated with generalized eta-squared ( ̂ 2). As a measure of the size of the fMR-adaptation effect, adaptation indices were computed [(different – same) / (different + same)] for faces and butterflies separately at each ROI. A 3x2 [Stimulation Condition (TMS to OFA, TMS to LO, Sham) x Image Category (Face or Butterfly)] omnibus repeated measures ANOVA was performed on the adaptation indices for each ROI. Our ROI analysis included face, object, and scene-preferential regions: the OFA, FFA, pSTS, LO, “parahippocampal place area” (PPA; Epstein and Kanwisher, 1998) and transverse occipital sulcus (TOS; Grill-Spector, 2003), which has also been called the “occipital place area” (OPA; Dilks et al., 2013).

2.2 Face-preferential ROIs 2.2.1 OFA In the right OFA, the experimental stimulation site, there were significant main effects of Image Category [F (1, 12) = 36.71, p < 0.001, ̂ 2 = 0.45] and Image Type [F (1, 12) = 81.21, p < 0.001, ̂ 2 = 0.16]. Consistent with its face preferential nature, activation was higher for images of faces than for butterflies, and there was an fMR-A effect as activation was higher for different than for same stimulus images. There was no main effect of Stimulation Condition [F (2, 24) = 1.56, p = 0.23]. The interaction between Stimulation Condition and Image Category approached significance [F (2, 24) = 2.81, p = 0.079, ̂ 2 = 0.006]. A priori FDR corrected pairwise comparisons revealed that activation for faces was lower with TMS to OFA compared to sham (p = 0.029), with no other significant comparisons (ps > 0.05).

7 In the left OFA, the hemisphere contralateral to stimulation, there were significant main effects of Image Category [F (1, 8) = 40.94, p < 0.001, ̂ 2 = 0.64] and Image Type [F (1, 8) = 65.34, p < 0.001, ̂ 2 = 0.25]. Consistent with its face preferential nature, activation was higher for images of faces than for butterflies, and there was an fMR-A effect as activation was higher for different than for same stimulus images. There was no main effect of Stimulation Condition [F (2, 16) = 2.49, p = 0.15], and no significant interactions (ps > 0.05), indicating no interhemispheric effect of the application of TMS to the right OFA in the left OFA.

2.2.2 FFA In the right FFA there were significant main effects of Image Category [F (1, 12) = 149.90, p < 0.001, ̂ 2 = 0.69] and Image Type [F (1, 12) = 99.53, p < 0.001, ̂ 2 = 0.30]. Consistent with its face preferential nature, activation was higher for images of faces than for butterflies, and there was an fMR-A effect as activation was higher for different than for same stimulus images. There was an interaction between Image Category and Image Type [F (1, 12) = 12.89, p = 0.0037, ̂ 2 = 0.018], such that the difference in activation between different and same stimuli was larger for faces than for butterflies. There was no main effect of Stimulation Condition [F (2, 24) = 1.87, p = 0.18]. The interaction between Stimulation Condition and Image Category approached significance [F (2, 24) = 2.94, p = 0.072, ̂ 2 = 0.026]. A priori FDR corrected pairwise comparisons revealed that activation for faces in the right FFA was lower with TMS to OFA compared to both sham (p = 0.0081) and TMS to LO (p = 0.013), with no other significant comparisons (ps > 0.05). In the left FFA, there were significant main effects of Image Category [F (1, 11) = 174.94, p < 0.001, ̂ 2 = 0.65] and Image Type [F (1, 11) = 113.98, p < 0.001, ̂ 2 = 0.32].

8 Consistent with its face preferential nature, activation was higher for images of faces than for butterflies, and there was an fMR-A effect as activation was higher for different than for same stimulus images. There was no main effect of Stimulation Condition [F (2, 22) = 2.19, p = 0.14], however interhemispheric effects of TMS were indicated by a significant interaction between Stimulation Condition and Image Category in the left FFA [F (2, 22) = 3.88, p = 0.036, ̂ 2 = 0.049]. FDR corrected pairwise comparisons revealed that activation for faces in the left FFA was lower with TMS to right OFA compared to sham (p = 0.024), with no other significant comparisons (ps > 0.05).

2.2.3 pSTS In the right pSTS, there was a significant main effect of Image Category [F (1, 12) = 158.16, p < 0.001, ̂ 2 = 0.56]. Consistent with its face preferential nature, activation was higher for face than for butterfly images. There was an interaction between Image Category and Image Type [F (1, 12) = 50.18, p < 0.001, ̂ 2 = 0.054]. Pairwise comparisons revealed the difference in activation between different and same stimuli was larger for butterflies than for faces. However, this larger difference in activation for butterflies than faces is the result of negative beta values and must be interpreted cautiously (Harel et al., 2002). There was also an interaction between Stimulation Condition and Image Type [F (2, 24) = 3.42, p = 0.049, ̂ 2 = 0.006]. FDR corrected pairwise comparisons revealed that activation for different stimuli was lower with TMS to OFA compared to sham (p = 0.02), with no other significant comparisons (ps > 0.05). The left pSTS could only be functionally localized in 5 out of 13 participants and was therefore omitted from analyses.

9 2.3 Non-face-preferential ROIs 2.3.1 LO In the right LO, the control stimulation site, there were significant main effects of Image Category [F (1, 12) = 66.79, p < 0.001, ̂ 2 = 0.51] and Image Type [F (1, 12) = 113.02, p < 0.001, ̂ 2 = 0.29]. Activation was higher for butterflies than faces, and for different than for same stimuli. There was an interaction between Image Category and Image Type [F (1, 12) = 11.10, p = 0.006, ̂ 2 = 0.040]. Pairwise comparisons revealed the difference in activation between different and same stimuli was larger for butterflies than for faces indicating a larger adaptation effect for the non-face category. There was also an interaction between Stimulation Condition and Image Category [F (2, 24) = 10.31, p < 0.001, ̂ 2 = 0.026]. FDR corrected pairwise comparisons revealed that butterfly activation in the right LO was lower with TMS to TMS to LO (p = 0.025) compared to sham, with no other significant comparisons (ps > 0.05). In the left LO, contralateral to stimulation, there were significant main effects of Image Category [F (1, 12) = 79.08, p < 0.001, ̂ 2 = 0.42] and Image Type [F (1, 12) = 154.96, p < 0.001, ̂ 2 = 0.18]. Activation was higher for butterflies than faces, and for different than same stimuli. There was an interaction between Image Category and Image Type [F (1, 12) = 19.25, p < 0.001, ̂ 2 = 0.034]. Pairwise comparisons revealed the difference in activation between different and same stimuli was larger for butterflies than for faces indicating a larger adaptation effect for the non-face category. There was no significant effect of Stimulation Condition [F (2, 24) = 1.46, p = 0.25], and no significant interactions (ps > 0.05), indicating no interhemispheric effect of the application of TMS to the right LO in the left LO.

2.3.2 TOS

10 In the right TOS, there were significant main effects of Image Category [F (1, 12) = 24.97 p < 0.001, ̂ 2 = 0.22] and Image Type [F (1, 12) = 36.49, p < 0.001, ̂ 2 = 0.05]. Activation was higher for butterflies than faces, and for different than same stimuli. There was an interaction between Image Category and Image Type [F (1, 12) = 10.96, p = 0.006, ̂ 2 = 0.018]. Pairwise comparisons revealed the difference in activation between different and same stimuli was larger for butterflies than for faces. There were no effects of TMS in the right TOS (ps > 0.05). Likewise, in the left TOS, there were significant main effects of Image Category [F (1, 12) = 21.37, p < 0.001, ̂ 2 = 0.13] and Image Type [F (1, 12) = 11.86, p = 0.004, ̂ 2 = 0.026]. Activation was higher for butterflies than faces, and for different than same stimuli. There was an interaction between Image Category and Image Type [F (1, 12) = 7.29, p = 0.019, ̂ 2 = 0.011]. Pairwise comparisons revealed the difference in activation between different and same stimuli was larger for butterflies than for faces. There were no effects of TMS in the left TOS (ps > 0.05). These data are consistent with both left and right TOS exhibiting more preferential activation for scenes and objects within scenes rather than faces.

2.3.3 PPA In the right PPA, there were significant main effects of Image Category [F (1, 12) = 36.96, p < 0.001, ̂ 2 = 0.27] and Image Type [F (1, 12) = 44.34, p < 0.001, ̂ 2 = 0.088]. Activation was higher for butterflies than faces, and for different than same stimuli. There was an interaction between Image Category and Image Type [F (1, 12) = 12.01, p < 0.001, ̂ 2 = 0.05]. Pairwise comparisons revealed the difference in activation between different and same stimuli

11 was larger for butterflies than for faces. There were no effects of TMS in the right PPA (ps > 0.05). Similarly, in the left PPA, there were significant main effects of Image Category [F (1, 12) = 18.45, p = 0.001, ̂ 2 = 0.17] and Image Type [F (1, 12) = 24.03, p < 0.001, ̂ 2 = 0.042]. Activation was higher for butterflies than faces, and for different than same stimuli. There was an interaction between Category and Type [F (1, 12) = 27.53, p < 0.001, ̂ 2 = 0.038]. Pairwise comparisons revealed the difference in activation between different and same stimuli was larger for butterflies than for faces. There were no effects of TMS in the left PPA (ps > 0.05). These data are consistent with both left and right PPA exhibiting more preferential activation for scenes and objects within scenes than faces.

2.4 Adaptation Indices The current study did not observe any effects of TMS on adaptation indices (ps > 0.05).

3. Discussion We showed that TMS to the right OFA selectively reduced face processing within this region but also within both the left and right FFA. Further, despite this attenuation of BOLD activity in the OFA and FFA, face adaptation effects persisted, suggesting that both regions continue to code face identity at some level. Together, this indicates that the integrity of the right OFA is necessary to maintain normal levels of face processing in both right and left FFA. While we observed interhemispheric effects of TMS to the right OFA in the left FFA, we observed no effects in the left OFA indicating that there may not be a direct connection from

12 right to left OFA. A number of models could explain the remote effects of TMS to right OFA. It could be that TMS to the right OFA selectively introduces neural noise in the right FFA, which then affects processing in the left FFA or alternatively, TMS to the right OFA could possibly affect both the right and left FFA simultaneously. What is clear is that activity in the OFA has direct consequences on these connected areas in the network. While the current TMS paradigm does not obliterate the ability of either the FFA or OFA to code face identity, it is important to observe that TMS to the OFA does selectively modulate face processing in both regions. Patients with acquired prosopagnosia who have lesions in the right OFA are unable to recognize the identity of faces despite showing face-selective activation at the right FFA and an intact ability to categorize faces (Dricot et al., 2008; Rossion et al., 2003; Schiltz et al., 2006; Steeves et al., 2006). The present data are consistent with data showing no release from fMRadaptation for face identity in prosopagnosia (Steeves et al., 2009) and further highlight the particular importance of the OFA in identity coding. While we did not observe changes in adaptation indices when the OFA was stimulated, other studies report behavioural impairments in face identity processing following TMS to the same area (e.g. Cohen Kadosh et al., 2011; Solomon-Harris et al., 2013). The lack of change in adaptation indices is consistent with an overall baseline shift in the level of signal following TMS. Again, it is important to be cognisant of the fact that TMS effects are often modest and more nuanced and not as extreme as a patient lesion. Consistent with our findings, intracranial stimulation to the right OFA in a patient with refractory epilepsy transiently impaired face identity discrimination (Jonas et al., 2014). TMS to the OFA also significantly affected BOLD signal for different stimulus images in the right pSTS, indicating direct or indirect connectivity between these regions. Another recent consecutive TMS-fMRI study found that neural noise in the OFA reduced activity in the FFA for

13 static and dynamic faces, while it reduced activity in the pSTS only for static faces. Further, TMS to the pSTS reduced activity in the pSTS only for dynamic faces, suggesting that static and dynamic face processing is achieved via dissociable cortical pathways (Pitcher et al., 2014). Furthermore, TMS to the right LO, an object-preferential region, significantly decreased activation for the non-face category of images, namely butterflies within this region. We did not observe interhemispheric effects of TMS to the right LO in the left LO in the present study. Previously we have observed mixed findings regarding interhemispheric effects of TMS to the LO (Rafique et al., 2015; Mullin et al., 2013), highlighting the important influence of different experimental paradigms. While TMS to both the OFA and LO decreased activation within these regions relative to sham stimulation for the preferred stimulus category, these differences were not significant relative to TMS at the other site. The lack of difference in activation for faces in the right OFA and butterflies in the right LO following TMS to these same stimulation sites suggests there could be a proximity effect with a possible spread of TMS between these two regions given the extended stimulation time. The three-dimensional (3D) distance between the OFA and LO in native space is 18.6 mm, a relatively close proximity allowing for adequate control of peripheral effects of TMS across stimulation sites, which are quite different depending on where stimulation occurs on the head (Duecker et al., 2013). Nonetheless, LO has served as an effective TMS control site for the OFA in our previous online behavioural study which used relatively short bursts of pulse trains (500 ms bursts; Solomon-Harris et al., 2013) and has also been used by others (e.g. Pitcher et al., 2011a; Pitcher et al., 2009; Pitcher et al., 2007). Previously we used consecutive TMS-fMRI by applying TMS to the left LO (rather than the right in the present study) but did not observe changes in face preferential activity in the left OFA or FFA (Mullin

14 and Steeves, 2013). In that study, however, TMS was applied for a shorter time period (15 min) compared to the current study (20 min), which could have potentially contributed to a spatial spread of TMS effect. It is also possible that TMS to LO and OFA had similar effects within LO and OFA if there is communication between these two regions. Dynamic causal modelling suggests that the LO may play a role in face processing as there are bidirectional connections between LO and both the OFA and FFA (Nagy et al., 2012). Further, diffusion tensor imaging has demonstrated anatomical connectivity between LO and the FFA (Kim et al., 2006) which is consistent with evidence of face discrimination in LO (Goesaert and Op de Beeck, 2013). Future research is necessary to disentangle the possible role of the LO in face processing. In short, we causally demonstrate that TMS to the right OFA selectively disrupts face processing in both the right and left FFA, indicating functional connectivity between these regions. We further demonstrate functional connectivity between the right OFA and pSTS, although this disruption was not selective to faces. We cannot determine whether connectivity between the OFA, FFA, and pSTS operates hierarchically or non-hierarchically based solely on these data. However, this research dovetails with mounting evidence from physiology (Jonas et al., 2014), patients with brain damage (Atkinson and Adolphs, 2011; Dricot et al., 2008; Rossion et al., 2003; Schiltz et al., 2006; Steeves et al., 2006; Steeves et al., 2009), and neuroimaging studies of the healthy brain (Goesaert and Op de Beeck, 2013; Goffaux et al., 2011; Jiang et al., 2015; Jiang et al., 2011; Rossion et al., 2011; Rossion et al., 2012) demonstrating the key role of the OFA for face recognition. Future work should continue to explore the specificity and connectivity of the face network, recognizing the limitations of purely hierarchical models in studying highly complex, interconnected systems.

15

4. Experimental Procedure 4.1 Participants Thirteen healthy volunteers (8 female, 11 right handed, mean age 29.1 years) participated in all three conditions of the experiment, as well as preliminary fMRI to localize the stimulation sites and ROIs. All participants had normal or corrected-to-normal vision and no known contraindications to TMS or fMRI. The York University Office of Research Ethics approved this study and participants were treated in accordance with the Declaration of Helsinki.

4.2 Data Acquisition and Preprocessing Structural and functional images were acquired using a 3 Tesla Siemens Magnetom Tim Trio magnetic resonance scanner at the York MRI Facility (Toronto, Canada) and the Siemens 32 channel head coil. High-resolution anatomical images were acquired with an MP-RAGE sequence (magnetization prepared rapid acquisition with gradient echo, in-plane resolution 1 x 1 mm, 176 sagittal slices, slice thickness = 1 mm, imaging matrix 256 × 256, FOV = 256 x 256 mm, TE = 2.52 ms, TR =1900 ms, flip angle = 9°, TI = 900 ms). Functional volumes were acquired with echo planar imaging (in-plane resolution 2.5 x 2.5 mm, slice thickness = 3 mm, 96 x 96 imaging matrix, FOV = 24 x 24 cm, 32 axial slices, TR = 2 s, TE = 30 ms, flip angle = 90°). Imaging analyses were performed using BrainVoyager QX software (Brain Innovation, Maastricht, NL). Functional data were subject to preprocessing steps including linear trend removal to exclude scanner-related signal drift, high-pass filtering to remove temporal frequencies lower than three cycles per run, and a correction for small interscan head movements

16 using a rigid body algorithm rotating and translating each functional volume in 3D space. Each participant’s functional images were coregistered with their anatomical images. The functional data were analysed using a general linear model.

4.3 Prestimulation fMRI Stimulation sites and ROIs for subsequent comparisons across TMS conditions were individually localized with fMRI in a pre-experimental session. Functional localizer scans used a block design and participants performed a one-back task to focus attention on the 3 categories of visual stimuli: colour images of faces, scenes and objects. Each run began and finished with a fixation cross for 16 s. Six repetitions of three 16 s blocks of the three categories of stimuli were presented in pseudorandom order. Each repetition was interleaved with 16 s of fixation. Each block contained 16 stimuli presented for 1 s each. Imaging data were collected over two functional runs (6 min, 52 s). Stimuli were presented with a rear-projection system (Avotec, Stuart, FL). Bilateral ROIs were localized in each participant by selecting the peak activation following the appropriate contrasts in a general linear model, and the extent of ROIs was restricted to relevant anatomical areas reported in the literature (FDR-corrected whole brain threshold of q < 0.05 and cluster threshold of 6). A linear balanced contrast of faces versus objects and scenes was used to localize face-preferential ROIs: the OFA (experimental TMS site), FFA, and pSTS. A linear balanced contrast of objects versus faces and scenes was used to localize the object-preferential area LO (control TMS site). A linear balanced contrast of scenes versus faces and objects was used to identify ROIs for scene-preferential regions: the PPA and TOS. Anatomical images from the localizer runs were transformed into Talairach space

17 (Talairach and Tournoux, 1988) and mean Talairach coordinates for the centre of each ROI were determined to be within the range of those reported in other studies (Table 1; e.g. Dricot et al., 2008; Ewbank et al., 2005; Mullin and Steeves, 2013; Steeves et al., 2009). For each ROI identified in the stimulated (right) hemisphere, its contralateral counterpart was also defined. Evaluation of contralateral ROIs allows the assessment of potential remote interhemispheric effects. However, the left pSTS could only be identified in 5/13 participants and was therefore omitted from analyses. Right hemisphere dominance in face processing, as well as smaller and less reliable activation for static faces in the pSTS, are consistent with the work of others (e.g. Bentin et al., 1996; Henson et al., 2003; Sergent et al., 1992).

4.4 TMS Functional Stereotaxy The functionally defined stimulation sites were targeted with Brainsight image-guided co-registration software and hardware (Rogue Research, Montréal, QC) using each individual’s MRI scans for each participant. Common reference points on both the MR images and the participant's head were selected to create a co-registration matrix. The spatial relationship between these reference points on the MR images and those on the participant's head were coregistered using a Polaris infrared marker system. The brain stimulation sites were individually selected by overlaying each participant’s activation map from the fMRI localizer onto a threedimensional reconstruction of the participant’s brain and scalp within the Brainsight software. Subsequently, image-guided TMS was achieved by monitoring, in real time, the location and orientation of the TMS coil and targeted brain stimulation site via infrared markers on the coil and the participant’s head.

18 4.5 Stimulation Parameters The experiment consisted of three stimulation conditions: (1) TMS to the right OFA, (2) TMS to the right LO, and (3) sham TMS to the right occipital lobe. The stimulation conditions were targeted in separate sessions on different days in counterbalanced order across participants. A Magstim Super Rapid2 stimulator and an air-cooled figure-of-eight coil with a diameter of 70 mm were used to deliver the stimulation pulses (Magstim, Whitland, UK). During stimulation, the coil was held tangential to the scalp surface with the handle pointed downward. For sham stimulation, the coil was positioned orthogonal to the scalp surface so that no pulse entered the brain. A low-frequency pulse (1 Hz) was delivered for 20 minutes (1200 pulses), thereby allowing at least 20 minutes of TMS-induced neural noise (Chen et al., 1997; Pascual-Leone et al., 1998; Robertson et al., 2003; Sandrini et al., 2011; Thut & Pascual-Leone, 2010) in which to assess potential effects on BOLD signal. The effects of low-frequency TMS have been demonstrated as strongest immediately following stimulation with a gradual decay to baseline (Eisenegger et al., 2008; Nyffeler et al., 2006), while other studies demonstrate a delay whereby effects are strongest a few minutes after the cessation of stimulation followed by a gradual decay to baseline (Chouinard et al., 2003; Johnson et al., 2007). Differences in the time-course of TMS effects are likely due to differences in the stimulation sites and tasks employed. The intensity was set at 60% of maximum stimulator output based on previous findings from our laboratory (Mullin and Steeves, 2011; Mullin and Steeves, 2013; Ganaden et al., 2013; Solomon-Harris et al., 2013) and others (Campana et al., 2002; Pitcher et al., 2011a; Pitcher et al., 2009; Pitcher et al., 2007; Silvanto et al., 2005). The frequency, intensity, and duration of the

19 TMS train were well within the safety limits of stimulation (Rossi et al., 2009; Wassermann, 1998). Earplugs were worn to dampen the noise from the coil discharge during TMS.

4.6 Poststimulation fMRI Immediately after each of the three TMS conditions, participants underwent functional neuroimaging. TMS was performed in the MRI control room in order to minimize the time between stimulation and neuroimaging, which was approximately 2 minutes. As soon as the participant was positioned in the scanner, the fMR-adaptation experiment was conducted first followed by structural image acquisition. The adaptation experiment used stimuli from previous fMRI studies of prosopagnosia (Dricot et al., 2008; Steeves et al., 2009) and was comprised of blocks of colour images of different identity faces, same identity faces, different identity butterflies, and same identity butterflies (Figure 1). To maintain attention, participants pressed a button to indicate when blocks switched between images of faces and butterflies (and vice versa). Each run began and finished with a fixation cross for 12 s. Eight repetitions of four 12 s blocks of the four categories of stimuli were presented in pseudorandom counterbalanced order. Each repetition was interleaved with 12 s of fixation. Each block contained 12 images presented for 800 ms followed by a 200 ms blank screen. Imaging data were collected over one functional run lasting 8 min 14 s. The independent prestimulation localizer ROIs were applied to the coregistered poststimulation data for each participant in order to measure the BOLD response after each TMS condition. Thresholds were held constant across pre- and poststimulation conditions for each ROI. The volume-of-interest analysis tool in BrainVoyager QX software was used to perform a

20 general linear model analysis (Brain Innovation, Maastricht, NL) and beta weights in the predefined ROIs were determined for each stimulation condition.

Table 1 Mean (SEM) Talairach coordinates for the functionally defined regions of interest (ROIs) ROI Number of Cluster Size Participants (mm3) x y z Face-preferential Right OFA 13 305 36 (2) -70 (3) -15 (2) Left OFA 9 311 -37 (3) -69 (3) -14 (2) Right FFA 13 386 36 (2) -47 (2) -19 (1) Left FFA 12 377 -37 (3) -48 (3) -20 (2) Right pSTS 13 339 49 (1) -46 (3) 9 (2) Left pSTS 5 296 -48 (1) -50 (1) 5 (1) Object-preferential Right LO 13 295 41 (2) -72 (2) -6 (1) Left LO 13 301 -42 (3) -71 (2) -6 (1) Scene-preferential Right TOS 13 349 33 (3) -82 (3) 16 (2) Left TOS 13 301 -32 (3) -82 (3) 14 (3) Right PPA 13 383 24 (3) -51 (4) -14 (3) Left PPA 13 326 -23 (4) -49 (3) -14 (2) Note. Each region was identified with a threshold of p < 0.05, FDR-corrected. OFA = occipital face area; FFA = fusiform face area; pSTS = posterior superior temporal sulcus; LO = lateral occipital area; TOS = transverse occipital sulcus; PPA = parahippocampal place area; SD = standard deviation; FDR = false discovery rate. Right and Left refer to the cerebral hemispheres.

21 Figure 1. A schematic of the face fMR-adaptation experiment which is similar to that used in (Steeves et al., 2009). Twelve-second blocks of 12 images (800 ms + 200 ms blank screen) depicting different faces, same faces, different butterflies, and same butterflies. Eight repetitions of the four stimulus category blocks were presented in pseudorandom counterbalanced order. Each repetition was interleaved with 12 s of fixation. Each run began and finished with a fixation cross for 12 s. Data were collected over one functional run lasting 8 min 14 s.

Figure 2. Effects of transcranial magnetic stimulation (TMS) in face-preferential regions. Error bars represent standard error of the mean. OFA = occipital face area; FFA = fusiform face area; pSTS = posterior superior temporal sulcus; n.s. = not significant; *p < 0.05; **p < 0.01 [false discovery rate (FDR) corrected p-values].

Figure 3. Effects of transcranial magnetic stimulation (TMS) in non-face-preferential regions. Error bars represent standard error of the mean. LO = lateral occipital area; TOS = transverse occipital sulcus; PPA = parahippocampal place area; *p < 0.05 [false discovery rate (FDR) corrected p-values].

Acknowledgements This research was supported by grants from the Natural Sciences and Engineering Research Council of Canada and the Canada Foundation for Innovation.

22 References Atkinson, A.P., & Adolphs, R. (2011). The neuropsychology of face perception: Beyond simple dissociations and functional selectivity. Philosophical Transactions of the Royal Society B, 366, 1726-1738. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological), 57(1), 289-300. Bentin, S., Allison, T., Puce, A., Perez, E., & McCarthy, G. (1996). Electrophysiological studies of face perception in humans. Journal of Cognitive Neuroscience, 8(6), 551-565. Campana, G., Cowey, A., & Walsh, V. (2002). Priming of motion direction and area V5/MT: A test of perceptual memory. Cerebral Cortex, 12, 663-669. Chen, R., Classen, J., Gerloff, C., Celnik, P., Wassermann, E.M., Hallett, M., & Cohen, L.G. (1997). Depression of motor cortex excitability by low-frequency transcranial magnetic stimulation. Neurology, 48, 1398-1403. Chouinard, P.A., Van Der Werf, Y.D., Leonard, G., & Paus, T. (2003). Modulating neural networks with transcranial magnetic stimulation applied over the dorsal premotor and primary motor cortices. Journal of Neurophysiology, 90, 1071-1083. Cohen Kadosh, C., Walsh, V., & Cohen Kadosh, R. (2011). Investigating face-property specific processing in the right OFA. Social Cognitive and Affective Neuroscience, 6, 58-65. Dilks, D.D., Julian, J.B., Paunov, A.M., & Kanwisher, N. (2013). The occipital place area is causally and selectively involved in scene perception. The Journal of Neuroscience, 33(4), 1331-1336.

23 Dricot, L., Sorger, B., Schiltz, C., Goebel, R., & Rossion, B. (2008). The roles of “face” and “non-face” areas during individual face perception: Evidence by fMRI adaptation in a brain-damaged prosopagnosic patient. NeuroImage, 40(1), 318-332. Duecker, F., de Graaf, T.A., Jacobs, C., & Sack, A.T. (2013). Time- and task-dependent nonneural effects of real and sham TMS. PLoS One, 8(9), e73813, 1-9. Eisenegger, C., Treyer, V., Fehr, E., & Knoch, D. (2008). Time-course of “off-line” prefrontal rTMS effects – a PET study. NeuroImage, 42, 379-384. Epstein, R., & Kanwisher, N. (1998). A cortical representation of the local visual environment. Nature, 392, 598-601. Ewbank, M.P., Schluppeck, D., & Andrews, T.J. (2005). fMR-adaptation reveals a distributed representation of inanimate objects and places in human visual cortex. NeuroImage, 28, 268-279. Felleman, D.J., & Van Essen, D.C. (1991). Distributed hierarchical processing in the primate cerebral cortex. Cerebral Cortex, 1(1), 1-47. Ganaden, R.E., Mullin, C.R., & Steeves, J.K.E. (2013). Transcranial magnetic stimulation to the transverse occipital sulcus affects scene but not object processing. Journal of Cognitive Neuroscience, 25(6), 961-968. Gauthier, I., Tarr, M.J., Moylan, J., Skudlarski, P., Gore, J.C., & Anderson, A.W. (2000). The fusiform “face area” is part of a network that processes faces at the individual level. Journal of Cognitive Neuroscience, 12(3), 495-504. Goesaert, E., & Op de Beeck, H.P. (2013). Representations of facial identity information in the ventral visual stream investigated with multivoxel pattern analyses. The Journal of Neuroscience, 33(19), 8549-8558.

24 Goffaux, V., Peters, J., Haubrechts, J., Schiltz, C., Jansma, B., & Goebel, R. (2011). From coarse to fine? Spatial and temporal dynamics of cortical face processing. Cerebral Cortex, 21(2), 467-476. Grill-Spector, K. (2003). The neural basis of object perception. Current Opinion in Neurobiology, 13(2), 159-166. Grill-Spector, K., & Malach, R. (2001). fMR-adaptation: A tool for studying the functional properties of human cortical neurons. Acta Psychologica, 107, 293-321. Harel, N., Lee, S.P., Nagaoka, T., Kim, D.S., & Kim, S.G. (2002). Origin of negative blood oxygenation level-dependent fMRI signals. Journal of Cerebral Blood Flow & Metabolism, 22, 908-917. Haxby, J.V., Hoffman, E.A., & Gobbini, M.I. (2000). The distributed human neural system for face perception. Trends in Cognitive Science, 4(6), 223-233. Henson, R.N., Goshen-Gottstein, Y., Ganel, T., Otten, L.J., Quayle, A., & Rugg, M.D. (2003). Electrophysiological and haemodynamic correlates of face perception, recognition and priming. Cerebral Cortex, 13, 793-805. Hochstein, S., & Ahissar, M. (2002). View from the top: Hierarchies and reverse hierarchies in the visual system. Neuron, 36, 791-804. Jiang, F., Badler, J.B., Righi, G., & Rossion, B. (2015). Category search speeds up face-selective fMRI responses in a non-hierarchical cortical face network. Cortex, 66, 69-80. Jiang, F., Dricot, L., Weber, J., Righi, G., Tarr, M.J., Goebel, R., & Rossion, B. (2011). Face categorization in visual scenes may start in a higher order area of the right fusiform gyrus: Evidence from dynamic visual stimulation in neuroimaging. Journal of Neurophysiology, 106, 2720-2736.

25 Johnson, J.A., Strafella, A.P., & Zatorre, R.J. (2007). The role of dorsolateral prefrontal cortex in bimodal divided attention: Two transcranial magnetic stimulation studies. Journal of Cognitive Neuroscience, 19(6), 907-920. Johnson, M.H. (2005). Subcortical face processing. Nature Reviews Neuroscience, 6, 766-774. Jonas, J., Rossion, B., Krieg, J., Koessler, L., Colnat-Coulbois, S., Vespignani, H., Jacques, C., Vignal, J.P., Brissart, H., & Maillard, L. (2014). Intracerebral electrical stimulation of a face-selective area in the right inferior occipital cortex impairs individual face discrimination. NeuroImage, 99, 487-497. Kanwisher, N. (2000). Domain specificity in face perception. Nature Neuroscience, 3(8), 759763. Kanwisher, N., McDermott, J., & Chun, M.M. (1997). The fusiform face area: A module in human extrastriate cortex specialized for face perception. Journal of Neuroscience, 17(11), 4302-4311. Kim, M., Ducros, M., Carlson, T., Ronen, I., He, S., Ugurbil, K., & Kim, D.S. (2006). Anatomical correlates of the functional organization in the human occipitotemporal cortex. Magnetic Resonance Imaging, 24, 583-590. Milner, A.D., Perrett, D.I., Johnston, R.S., Benson, P.J., Jordan, T.R., Heeley, D.W., Bettucci, D., Mortara, F., Mutani, R., Terazzi, E., & Davidson, D.L.W. (1991). Perception and action in ‘visual form agnosia’. Brain, 114(1), 405-428. Mullin, C.R., & Steeves, J.K.E. (2011). TMS to the lateral occipital cortex disrupts object processing but facilitates scene processing. Journal of Cognitive Neuroscience, 23(12), 4174-4184.

26 Mullin, C. R., & Steeves, J. K.E. (2013). Consecutive TMS-fMRI reveals an inverse relationship in BOLD signal between object and scene processing. The Journal of Neuroscience, 33(49), 19243-19249. Nagy, K., Greenlee, M.W., & Kovács, G. (2012). The lateral occipital cortex in the face perception network: An effective connectivity study. Frontiers in Psychology, 3, article 141, 1-12. Nyffeler, T., Wurtz, P., Luscher, H.R., Hess, C.W., Senn, W., Pflugshaupt, T., von Wartburg, R., Luthi, M., & Muri, R.M. (2006). Repetitive TMS over the human oculomotor cortex: Comparison of 1-Hz and theta burst stimulation. Neuroscience Letters, 409, 57-60. Pascual-Leone, A. Tormos, J.M., Keenan, J., Tarazona, F., Cañete, C., & Catalá, M.D. (1998). Study and modulation of human cortical excitability with transcranial magnetic stimulation. Journal of Clinical Neurophysiology, 15(4), 333-343. Pitcher, D., Charles, L., Devlin, J.T., Walsh, V., & Duchaine, B. (2009). Triple dissociation of faces, bodies and objects in the extrastriate cortex. Current Biology, 19(4), 319-324. Pitcher, D., Duchaine, B., & Walsh, V. (2014). Combined TMS and fMRI reveal dissociable cortical pathways for dynamic and static face perception. Current Biology, 24, 2066-2070. Pitcher, D., Duchaine, B., Walsh, V., Yovel, G., & Kanwisher, N. (2011). The role of lateral occipital face and object areas in the face inversion effect. Neuropsychologia, 49, 34483453. Pitcher, D., Walsh, V., & Duchaine, B. (2011). The role of the occipital face area in the cortical face perception network. Experimental Brain Research, 209(4), 481-493. Pitcher, D., Walsh, V., Yovel, G., & Duchaine, B. (2007). TMS Evidence for the involvement of the right occipital face area in early face processing. Current Biology, 17, 1568-1573.

27 Puce, A., Allison, T., Bentin, S., Gore, J.C., & McCarthy, G. (1998). Temporal cortex activation in humans viewing eye and mouth movements. The Journal of Neuroscience, 18(6), 21882199. Rafique, S.A., Solomon-Harris, L.M., & Steeves, J.K.E. (2015). TMS to object cortex affects both object and scene remote networks while TMS to scene cortex only affects scene networks. Neuropsychologia, 79(A), 86–96. Rhodes, G., Michie, P.T., Hughes, M.E., & Byatt, G. (2009). The fusiform face area and occipital face area show sensitivity to spatial relations in faces. European Journal of Neuroscience, 30, 721-733. Robertson, E. M., Theoret, H., & Pascual-Leone, A. (2003). Studies in cognition: The problems solved and created by transcranial magnetic stimulation. Journal of Cognitive Neuroscience, 15(7), 948-960. Rossi, S., Hallett, M., Rossini, P.M., Pascual-Leone, A., & the safety of TMS consensus group. (2009). Safety, ethical considerations, and application guidelines for the use of transcranial magnetic stimulation in clinical practice and research. Clinical Neurophysiology, 120, 2008-2039. Rossion, B. (2008). Constraining the cortical face network by neuroimaging studies of acquired prosopagnosia. NeuroImage, 40, 423-426. Rossion, B., Caldara, R., Seghier, M., Schuller, A.M., Lazeyras, F., & Mayer, E. (2003). A network of occipito-temporal face-sensitive areas besides the right middle fusiform gyrus is necessary for normal face processing. Brain, 126, 2381-2395.

28 Rossion, B., Dricot, L., Goebel, R., & Busigny, T. (2011). Holistic face categorization in higher order visual areas of the normal and prosopagnosic brain: Toward a non-hierarchical view of face perception. Frontiers in Human Neuroscience, 4, article 225, 1-30. Rossion, B., Hanseeuw, B., & Dricot, L. (2012). Defining face perception areas in the human brain: A large-scale factorial fMRI face localizer analysis. Brain and Cognition, 79, 138157. Sandrini, M., Umiltà, C., & Rusconi, E. (2011). The use of transcranial magnetic stimulation in cognitive neuroscience: A new synthesis of methodological issues. Neuroscience and Biobehavioral Reviews, 35, 516-536. Schiltz, C., Sorger, B., Caldara, R., Ahmed, F., Mayer, E., Goebel, R., & Rossion, B. (2006). Impaired face discrimination in acquired prosopagnosia is associated with abnormal response to individual faces in the right middle fusiform gyrus. Cerebral Cortex, 16, 574586. Sergent, J., Ohta, S., & MacDonald, B. (1992). Functional neuroanatomy of face and object processing: A positron emission tomography study. Brain, 115(1), 15-36. Silvanto, J., Lavie, N., & Walsh, V. (2005). Double dissociation of V1 and V5/MT activity in visual awareness. Cerebral Cortex, 15(11), 1736-1741. Solomon-Harris, L.M., Mullin, C.R., & Steeves, J.K.E. (2013). TMS to the “occipital face area” affects recognition but not categorization of faces. Brain and Cognition, 83(3), 245-251. Steeves, J.K.E., Culham, J.C., Duchaine, B.C., Pratesi, C.C., Valyear, K.F., Schindler, I., Humphrey, G.K., Milner, A.D., & Goodale, M.A. (2006). The fusiform face area is not sufficient for face recognition: Evidence from a patient with dense prosopagnosia and no occipital face area. Neuropsychologia, 44, 594-609.

29 Steeves, J., Dricot, L., Goltz, H.C., Sorger, B., Peters, J., Milner, A.D., Goodale, M.A., Goebel, R., & Rossion, B. (2009). Abnormal face identity coding in the middle fusiform gyrus of two brain-damaged prosopagnosic patients. Neuropsychologia, 47(12), 2584-2592. Talairach, J. & Tournoux, P. (1988). Co-planar stereotaxic atlas of the human brain. New York, NY: Thieme. Thut, G., & Pascual-Leone, A. (2010). A review of combined TMS-EEG studies to characterize lasting effects of repetitive TMS and assess their usefulness in cognitive and clinical neuroscience. Brain Topography, 22, 219-232. Ungerleider, L.G., & Mishkin, M. (1982). Two cortical visual systems. In D.J. Ingle, M.A. Goodale, and R.J.W. Mansfield (Eds.), Analysis of Visual Behavior (pp. 549-586). Cambridge, MA: The MIT Press. Wassermann, E.M. (1998). Risk and safety of repetitive transcranial magnetic stimulation: Report and suggested guidelines from the international workshop on the safety of repetitive transcranial magnetic stimulation, June 5-7, 1996. Electroencephalography and Clinical Neurophysiology, 108, 1-16. Xu, X., & Biederman, I. (2010). Loci of the release from fMRI adaptation for changes in facial expression, identity, and viewpoint. Journal of Vision, 10(14), 1-13.

30 Highlights     

We used TMS to introduce neural noise in the right OFA in healthy subjects. We then used fMRI to measure TMS-related changes in BOLD signal. TMS to OFA reduces BOLD signal for faces in the OFA and bilateral FFA. TMS to OFA also reduces BOLD signal for different identities in the right pSTS. This causally demonstrates that the OFA modulates activity across the face network.

31

32

33