European Journal of Neuroscience, Vol. 42, pp. 1905–1911, 2015
doi:10.1111/ejn.12950
COGNITIVE NEUROSCIENCE
Cooperative processing in primary somatosensory cortex and posterior parietal cortex during tactile working memory Yixuan Ku,1,2,3,* Di Zhao,1,* Mark Bodner4 and Yong-Di Zhou3,5 1
The Key Lab of Brain Functional Genomics, MOE & STCSM, Institute of Cognitive Neuroscience, 3663, North Zhongshan Road, School of Psychology and Cognitive Science, East China Normal University, Shanghai 200062, China 2 Departments of Neurology, Physiology and Psychiatry, University of California, San Francisco, CA, USA 3 NYU-ECNU Institute of Brain and Cognitive Science, NYU Shanghai and Collaborative Innovation Center for Brain Science, Shanghai, China 4 MIND Research Institute, Irvine, CA, USA 5 Krieger Mind/Brain Institute, Johns Hopkins University, 3400 N. Charles Street, 338 Krieger Hall, Baltimore, MA 21218, USA Keywords: cooperative processing, posterior parietal cortex, primary somatosensory cortex, single-pulse transcranial magnetic stimulation, tactile working memory
Abstract In the present study, causal roles of both the primary somatosensory cortex (SI) and the posterior parietal cortex (PPC) were investigated in a tactile unimodal working memory (WM) task. Individual magnetic resonance imaging-based single-pulse transcranial magnetic stimulation (spTMS) was applied, respectively, to the left SI (ipsilateral to tactile stimuli), right SI (contralateral to tactile stimuli) and right PPC (contralateral to tactile stimuli), while human participants were performing a tactile-tactile unimodal delayed matching-to-sample task. The time points of spTMS were 300, 600 and 900 ms after the onset of the tactile sample stimulus (duration: 200 ms). Compared with ipsilateral SI, application of spTMS over either contralateral SI or contralateral PPC at those time points significantly impaired the accuracy of task performance. Meanwhile, the deterioration in accuracy did not vary with the stimulating time points. Together, these results indicate that the tactile information is processed cooperatively by SI and PPC in the same hemisphere, starting from the early delay of the tactile unimodal WM task. This pattern of processing of tactile information is different from the pattern in tactile-visual cross-modal WM. In a tactile-visual cross-modal WM task, SI and PPC contribute to the processing sequentially, suggesting a process of sensory information transfer during the early delay between modalities.
Introduction Working memory (WM) refers to cognitive processes that manipulate and maintain sensory information for a short term to guide goal-directed behavioral actions (Baddeley, 2003). Persistent activation of neurons in association cortices [e.g. the prefrontal cortex (PFC) and the posterior parietal cortex (PPC)] and sensory cortices during a delay period of a WM task has been indicated by studies to represent the neural mechanisms underlying those maintaining and manipulating processes (see reviews in Pasternak & Greenlee, 2005; Curtis & Lee, 2010). However, the temporal dependence of these areas in WM processes still remains unclear. Hierarchical (Fuster, 2001) and parallel (Ballard, 1986; GoldmanRakic, 1988) models have been proposed for WM information
Correspondence: Dr Yixuan Ku, 1The Key Lab of Brain Functional Genomics, as above. E-mail: [email protected] Dr Yong-Di Zhou, 5Krieger Mind/Brain Institute, as above. E-mail: [email protected] *Y.K. and D.Z. contributed equally to this work. Received 10 February 2015, revised 25 April 2015, accepted 13 May 2015
processing, and both gain a plethora of supports (Fuster & Bressler, 2012). Nevertheless, in the somatosensory domain, the PPC is more frequently recognised to be a higher stage compared with the primary somatosensory cortex (SI; Iwamura, 1998; Fuster, 2001; Fuster & Bressler, 2012). However, to date there is still a lack of evidence for the hierarchical relationship between PPC and SI in tactile WM. Posterior parietal cortex has been reported to play an important role in tactile WM. Earlier neurophysiological work with primates has revealed that neurons in PPC show elevated and selective activities towards tactile stimuli through the delay period (Koch & Fuster, 1989). Recent neuroimaging studies with both positron emission tomography and functional magnetic resonance imaging (MRI) have shown that PPC is activated during WM (Klingberg et al., 1996; Preuschhof et al., 2006; Kaas et al., 2013). Meanwhile, recent neurophysiological data have suggested important roles of lower-level sensory cortices [SI and secondary somatosensory cortex (SII)] in tactile WM (Zhou & Fuster, 1996; Salinas et al., 2000; Romo & Salinas, 2003). The causal role of SI in maintaining vibrotactile memory traces has been reported in a human transcranial magnetic stimulation (TMS) study (Harris et al., 2002). Although the role
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd
1906 Y. Ku et al. played by SI in tactile WM has been welcomed by many researchers, there are still studies that do not support the finding (Hannula et al., 2010). In this present study, PPC was further recruited as a stimulating site, and the roles of SI and PPC were jointly assessed in a tactile unimodal delayed matching-to-sample task. Single-pulse TMS (spTMS) was used to stimulate SI and PPC at distinct temporal stages in the task. This stimulation is non-invasive and widely used to test the casual influence of brain activity over behavioral performance, especially regarding the time course of the involvement of a specific cortical area in a given cognitive process (M€ uri et al., 1996; Pascual-Leone et al., 2000). By using spTMS, sequential roles of SI and PPC in a tactilevisual cross-modal WM task (Ku et al., 2015) have recently been shown. Here, whether such a sequential relationship between SI and PPC would also exist in tactile unimodal WM was tested.
Materials and methods Participants Fifteen individuals participated in the experiment (11 females and four males, mean age = 22.7 years). None of the participants had neurological, psychiatric or other relevant medical problems. They did not have any contraindication to TMS. They were given written consent forms detailing the potential risks of the TMS experiment and signed to participate. After the experiment, they were paid for compensation. Twelve participants (out of 15) have participated in the cross-modal experiment in the previous study (Ku et al., 2015). All experimental protocols were approved by the Institutional Review Board of East China Normal University. Experimental paradigm A participant sat in a comfortable chair in front of a 17-inch CRT monitor (IBM C220P CRT; resolution ratio = 800 9 600 pixels; monitor refresh interval = 16.66 ms), with a chin support situated 0.87 m away from the monitor. The experiment consisted of a tactile-tactile unimodal delayed matching-to-sample task. In this task, a complete trial included a series of events as illustrated in Fig. 1, starting with a gray fixation-cross (lasting 1000–1500 ms) at the center of the monitor. The participant was asked to focus his/her attention on the gray cross. When the gray cross turned red, a tactile vibration (S-1, 200 ms duration) was applied to the left index finger of the participant. Frequencies of the vibration were 30, 40, 60 or
100 Hz, based on the equal sensation contours for vibration (Goff, 1967). A delay interval of 1000 ms followed S-1 when the central cross turned back to gray (offset of S-1). During the delay, the participant was instructed to focus his/her attention on the central cross again, and maintain in his/her mind the vibration frequency of S-1. The delay ended with the onset of a second stimulus (S-2, 200 ms duration) that was also a tactile vibration from the same vibration sets of S-1. The cross at the center of the monitor screen turned red during the presence of S-2. The participant reported whether S-2 matched S-1 by pressing one of two buttons as quickly and accurately as possible. The button-assignment (e.g. the right index finger for match, the right middle finger for non-match) was counterbalanced across participants. Both the participant’s choice and response time (RT; the span from the offset of S-2 to the choice) were recorded. A complete trial ended with the choice. There was an interval between two trials, which was randomly set between 3 and 4 s. Each participant performed three training blocks (with feedback of performance information, a correct or an incorrect response) to learn the task, and subsequently performed six experimental TMS blocks (without feedback). Each of these blocks consisted of 48 trials (24 ‘match’ trials and 24 ‘non-match’ trials, randomly arranged). Tactile stimulator A permanent magnetic vibrator (LDS V101 vibrator; probe diameter, 6.4 mm) driven by a LDS PA25E Power Amplifier (Br€ uel & Kjær Sound & Vibration. Measurement A/S, Denmark) was used as a tactile stimulator. The amplitude of a vibration at different frequencies was restricted to the same level (vertical displacement, 0.049 inches). The participant rested his/her left index finger at a fixed location on the vibrator. TMS protocol A similar TMS protocol was operated as in the previous study (Ku et al., 2015). TMS was delivered with a Magstim Rapid2 stimulator and a 70-mm figure-eight-shaped coil (The Magstim Company, Whitland, UK). High-resolution anatomical T1-weighted MRIs were acquired with a 3T Siemens Trio MRI scanner for each participant (TR = 2530 ms, TE = 2.340 ms, flip angle = 35°, FOV = 256 mm, 170 slices, 1 mm thickness) at the ECNU MRI Research Center. These images were then imported into the BrainSightTM neuronavigation software (BrainSight 2.0, Rogue Research, Montreal, Canada)
Fig. 1. Experimental paradigm. In this tactile-tactile unimodal delayed matching-to-sample task, the participant is asked to report whether the second stimulus matches the first stimulus in each trial. Single-pulse transcranial magnetic stimulation (spTMS) is delivered during the delay period at 300/600/900 ms after the onset of the first stimulus. © 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
SI and PPC in tactile WM processing 1907 to allow for stereotaxic registration of the TMS coil with the participant’s brain areas. The resting motor threshold that represented the cortical excitability threshold for each participant was measured by the minimal intensity necessary to elicit a visible movement of the relaxed abductor pollicis brevis muscle of the right hand in five out of 10 TMS pulses applied over the motor hand area in the left hemisphere (Kammer et al., 2001). The strength of TMS for each participant was then calculated as 110% of the motor threshold (mean: 67.6% maximum machine output; standard deviation: 8.2%). The motor hand area was also marked for each participant within the Brainsight software. The left SI (ipsilateral SI to the tactile stimuli, iSI) was localised as a point 2 cm posterior to the motor spot (Meehan et al., 2011). The mean Talairach coordinates of iSI across participants were as follows: x = 34 mm, y = 36 mm, z = 51 mm (see individual locations in Fig. 2, yellow balls). The contralateral SI (cSI) was localized as a mirror point of iSI towards the sagittal plane of the brain. The mean Talairach coordinates of cSI were: x = 34 mm, y = 38 mm,
z = 51 mm (see individual locations in Fig. 2, red balls). The contralateral PPC (cPPC) was localized as the spot about 1 cm posterior and 2 cm lateral from cSI (Pasalar et al., 2010). The mean Talairach coordinates of cPPC were: x = 45 mm, y = 46 mm, z = 50 mm (see individual locations in Fig. 2, blue balls), which was located in the right inferior parietal lobule. The TMS coil was held to the head of the participant with a custom coil holder, and was orientated about 45° clockwise away from the sagittal line at all stimulating sites and tangentially to the scalp. spTMS was then delivered, respectively, over the iSI, cSI and cPPC (two experimental blocks for each stimulating brain site) when the participant performed the task (the order of blocks assigned randomly for each participant). Three different stimulating time points (STPs: 300, 600 and 900 ms after the onset of S-1) for spTMS were the same as in the previous study (Ku et al., 2015). Within each block, 48 trials were equally divided into three groups in line with the STPs and, in each group, the four vibration frequencies were arranged pseudo-randomly, with each of them having four trials. All 48 trials in a single block were shuffled randomly for each participant. Paradigms and TMS were programmed using Matlab Psychtoolbox-3 (www.psychtoolbox.org) based on Matlab R2010b (MathWorks, MA, USA). White noise To attenuate influences from the noise that accompanied vibrations and TMS, earplugs were used and, in addition, white noise (80 dB) was generated using Adobe Audition 3.0 (Adobe Systems, USA) and delivered during the whole experiment through two loudspeakers placed on two sides of the CRT, respectively. Statistical analysis
Fig. 2. The stimulating sites across participants [yellow balls, ipsilateral primary somatosensory cortex (iSI); red balls, contralateral (cSI); blue balls, contralateral posterior parietal cortex (cPPC)].
A
The accuracy of task performance was defined as the percentage of correct responses in each condition, and the RT was defined as the interval between the offset of S-2 and the subject’s correct response. They were analysed by a two-way repeated-measures analysis of variance (RM-ANOVA), with LOC (stimulating sites: iSI, cSI and cPPC) and STP as within-subject factors. A two-sided paired Student’s t-test was performed whenever there were significant main effects or interactions. Multiple comparisons were corrected using the false discovery rate (FDR) method (Benjamini & Yekutieli, 2001).
B
Fig. 3. (A) Accuracy and (B) response time (RT) of behavioral performance across conditions. Dashed lines represent the mean of baseline performance in the last training block. The gray shaded area indicates the standard error of baseline performance. Error bars denote SEs. © 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
1908 Y. Ku et al. The participant’s performance in the third training block was extracted as a baseline without stimulation. TMS effects on performance were also compared between situations of TMS stimulation and baseline by using the paired t-test. Data from the 12 participants who participated in both unimodal and cross-modal studies were used to compare their performance between the studies. The baseline accuracy of performance in each study was used to normalise the accuracy data. The spTMS cost was defined as the percentage of drop-off from the baseline accuracy to the accuracy when spTMS was applied (as illustrated in the following equation).
CostAccuracy ¼
Accuracybaseline AccuracyspTMS 100% Accuracybaseline
All statistical analyses were performed with Statistica 6.0 (StatSoft, Tulsa, OK, USA).
Results After three blocks of training, participants performed the tasks with accuracy above 80% in the last training block (Fig. 3A). The task performance under all conditions was illustrated in Fig. 3 (accuracy, Fig. 3A; RT, Fig. 3B). RM-ANOVA for accuracy revealed a significant main effect of LOC (F2,28 = 9.25, P = 0.0008, gp2 = 0.398), but neither a significant main effect of STP (F2,28 = 0.529, P = 0.595, gp2 = 0.036) nor an interaction between LOC and STP (F4,56 = 0.233, P = 0.919, gp2 = 0.016) was observed. Post hoc analysis for LOC revealed that compared with iSI, stimulation on cSI or cPPC significantly (both FDR corrected, P < 0.05) decreased the accuracy of performance (Fig. 4). Further comparisons against baseline showed that stimulation on cSI or cPPC also significantly (both FDR corrected, P < 0.01) deteriorated the accuracy (Fig. 4), but stimulation on iSI did not (t14 = 1.39, P = 0.186).
RM-ANOVA for RT revealed neither a significant main effect of the factors, nor an interaction between them (Fig. 3B). To compare the results in the present unimodal study with those in the previous cross-modal study (Ku et al., 2015), first it was assessed whether there was any significant difference in baseline performance between the unimodal and cross-modal tasks. The twosided paired t-test revealed that there was no significant difference in baseline accuracy between those two tasks (t11 = 1.23, P = 0.24), indicating that the task difficulty was similar between them. However, baseline RT in the cross-modal task (mean SEM: 661 51 ms) was significantly longer than in the unimodal task (509 45 ms; t11 = 2.86, P = 0.01), which was likely due to a distinct sensory modality that S-2 (visual stimuli) had in the crossmodal task. Therefore, this study mainly focused on accuracy in conducting a comparison between unimodal and cross-modal task performance. spTMS showed sequentially deteriorative effects on cSI at STP 300 ms and on cPPC at 600 ms in the cross-modal task (Ku et al., 2015), but in the unimodal task (present study), spTMS showed deteriorative effects on cSI and cPPC across all STPs (Fig. 5). Thus, the cost of spTMS in these two conditions was assessed with a three-way RM-ANOVA, using TASK (cross-modal and unimodal), LOC (cSI and cPPC) and STP (300 and 600 ms) as within-subject factors. A marginal TASK*LOC*STP (F1,11 = 4.45, P = 0.0587, gp2 = 0.288) interaction was spotted. Further two-way RM-ANOVA with TASK and LOC revealed that, at STP 300 ms, there was a significant main effect of LOC (F1,11 = 8.52, P = 0.014, gp2 = 0.437) and a significant interaction between TASK and LOC (F1,11 = 14.4, P = 0.003, gp2 = 0.567), while at STP 600 ms, neither significant main effect nor interaction was observed. The individual spTMS cost at STP 300 ms was shown in Fig. 6A and B. Costs were higher under cSI stimulation than under cPPC stimulation in the cross-modal task (Fig. 6A), while they varied in the unimodal task (Fig. 6B). Cost differences between cSI and cPPC (cSI – cPPC) of all participants were compared between the two tasks, a significantly higher average cost was observed in the crossmodal task (t11 = 3.797, P = 0.003, as in Fig. 6C).
Discussion The main finding of the present study is that SI and PPC in the same hemisphere are both essential in tactile information processing in the delay period of a tactile unimodal WM task. Those two cortical areas work together in the processing, starting from the early delay (not later than 300 ms after the onset of tactile sample stimuli) through a period of at least 600 ms. Further, based on detailed comparisons of the current results of TMS effects received from the tactile unimodal WM task with the previous findings of TMS effects received from a tactile-visual cross-modal WM task (Ku et al., 2015), the authors believe that it is a reasonable assumption that the cross-modal information transfer occurs during the period between 300 and 600 ms after the onset of the sample stimulus (a duration of 200 ms). Involvement of SI and PPC during tactile WM
Fig. 4. The main effect of stimulating location. Dashed lines represent the mean of baseline performance in the last training block. The gray shaded area indicates the standard error of baseline performance. [false discovery rate (FDR) corrected, *P < 0.05, **P < 0.01; error bars denote SEs].
In most experiments of spTMS on frontal and parietal areas to date, the stimulation disturbed brain activities and downgraded performance in WM tasks (M€ uri et al., 1996; Mull & Seyal, 2001; Oliveri et al., 2001; Harris et al., 2002; D’Ardenne et al., 2012). Consistent with those studies, manipulation of TMS in the current study shows significant impairment of participants’ performance at all STPs,
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
SI and PPC in tactile WM processing 1909 A
B
Fig. 5. Comparisons of stimulating costs between cross-modal and unimodal tasks. Accuracy cost of (A) contralateral posterior parietal cortex (cPPC) vs. (B) contralateral primary somatosensory cortex (cSI).
A
B
C
Fig. 6. Comparisons of the stimulating cost between cross-modal and unimodal tasks at stimulating time point (STP) 300 ms. Scatter plot across participants in (A) the cross-modal task and (B) the unimodal task [accuracy cost of contralateral posterior parietal cortex (cPPC) vs. contralateral primary somatosensory cortex (cSI)]. (C) Differences in cost between cSI and cPPC across participants in cross-modal and unimodal tasks. Each circle represents an individual participant from whom data were received from both tasks in the study [data from the additional three participants for the unimodal study are only shown in (B) as squares].
indicating that both cSI and cPPC are task-related brain areas and play causal roles for tactile WM processing. The results are reminiscent of those neuroimaging findings on the involvement of PPC during tactile WM maintenance (Klingberg et al., 1996; Preuschhof et al., 2006; Savini et al., 2012; Kaas et al., 2013). The effects on performance are not likely due to clicking sounds induced by TMS because, on all stimulating sites, noise produced by stimulation is essentially at the same level. In a previous empirical study, Harris et al. (2002) used spTMS to explore the causal role of cSI in a tactile WM task. spTMS on cSI exhibited more severe decline of performance than on iSI, when stimulation was applied at 300 ms and 600 ms after the offset of tactile stimuli, but not at 900 and 1200 ms (Harris et al., 2002). The current results of cSI stimulation coincide with their findings at STPs 300 and 600 ms. However, a significant effect with spTMS on cSI at STP 900 ms has also been observed. This result may not contradict with Harris and colleagues’ findings, as the duration of tactile stimuli used in the current study (200 ms) is much shorter than theirs (1000 ms). In fact, the STPs in the current study all lie within the duration of their stimuli. Cooperative processing in SI and PPC during unimodal WM It has long been debated for somatosensory information processing at earlier sensory stages, whether serial or parallel processing dominates within SI and SII (Pons et al., 1987, 1992; Zhang et al., 1996; Iwamura, 1998; Karhu & Tesche, 1999; Ploner et al., 1999;
Inui et al., 2004; Liang et al., 2011; Kalberlah et al., 2013). Meanwhile, PPC has been widely accepted as a later stage of somatosensory processing and WM (Iwamura, 1998; Fuster, 2001; Fuster & Bressler, 2012). In the present study, a non-significant main effect of STP in both cSI and cPPC has been observed, suggesting that SI and PPC remain cooperative in processing sensory information across the delay period (at least starting at 300 ms after the onset of the sensory cue). Compared with the sequential processing pattern in the previous cross-modal WM study (Ku et al., 2015), the processing pattern here in the unimodal WM indicates distinct neural mechanisms. The unimodal WM does not demand an information transfer from one sensory modality to another, but requires maintenance of information in the same modality across the delay period and, therefore, a lag of 300 ms is not observed in TMS effects on PPC in the unimodal task. Apparently, this 300 ms delay in the previous cross-modal study is very likely for the information transfer between modalities. Further, stimulation of either cSI or cPPC shows TMS effects on task performance at all STPs, which indicates that to properly keep tactile information during the delay period, both SI and PPC are required. That is, sensory information is represented in both areas, each of which is critical in WM. These results further support a distributed WM network proposal (Zhou et al., 2007; Fuster & Bressler, 2012). Future study to explore the temporal profile of the earlier stage of sensory processing (during the presentation of stimuli) will be of interest, and future research with an extension of the delay period to
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
1910 Y. Ku et al. test if the conclusion of cooperative processing in SI and PPC during a long delay still holds will also be interesting. Primary sensory cortices in WM maintenance Compared with studies of sustained, elevated delay activity in the PFC (Fuster & Alexander, 1971; Goldman-Rakic, 1995; Miller & Cohen, 2001; Curtis & Lee, 2010), fewer studies have revealed sustained delay activity in primary sensory cortices in WM tasks (Zhou & Fuster, 1996, 2000; Super et al., 2001; Romo & Salinas, 2003; Bigelow et al., 2014). However, recent advances in neuroimaging have presented the possibility to accurately decode the cortical representation of sensory information during the WM (see reviews in Haynes & Rees, 2006; Davis & Poldrack, 2013), and thus revealed with multivariate pattern analysis content-specific representations in primary visual regions during WM (Harrison & Tong, 2009; Serences et al., 2009). Harrison and Tong have shown that even if the overall delay activity is low in human visual cortices, orientations held in WM can still be clearly decoded from activity patterns (Harrison & Tong, 2009). A recent study by Ester et al. (2013) has further revealed that information decoded in the primary sensory cortex is restricted for those remembered features. All those results indicate that primary sensory cortices are also involved in maintenance of sensory information during WM (Pasternak & Greenlee, 2005), and the current results provide further evidence for such involvement.
points did not reveal different effects between cSI and cPPC. Together, these results indicate cooperative processing in SI and PPC starting at an early stage during tactile unimodal WM. Comparisons in the TMS effects between unimodal and cross-modal tasks lead to a reasonable assumption that the cross-modal information transfer likely occurs during the period between 300 and 600 ms.
Acknowledgements The authors thank Hongyuan Li for his help on the interface of the tactile vibrator, and Liping Wang for his comments on the earlier proposals. This work was supported by National Key Fundamental Research (973) Program (2013CB329501), National Natural Science Foundation of China (31100742), Shanghai Committee of Science and Technology (15ZR1410600), and a research fund from the M.I.N.D. Research Institute, California.
Abbreviations cPPC, contralateral posterior parietal cortex; cSI, contralateral primary somatosensory cortex; FDR, false discovery rate; iSI, ipsilateral primary somatosensory cortex; MRI, magnetic resonance imaging; PFC, prefrontal cortex; PPC, posterior parietal cortex; RT, response time; SI, primary somatosensory cortex; SII, secondary somatosensory cortex; spTMS, singlepulse transcranial magnetic stimulation; STP, stimulating time point; TMS, transcranial magnetic stimulation; WM, working memory.
Conflict of interest Primary somatosensory cortices in cross-modal information transfer
None to declare.
It is notable that SI collaborates with PPC in different ways between the unimodal and cross-modal tasks, especially at 300 ms after the onset of the sample stimulus. In the cross-modal task, stimulation of cSI always disrupts performance more severe than that of cPPC (Fig. 6A), while in the unimodal task, stimulation of those two areas shows compatible degradation of performance (Fig. 6B). When contrasted with cPPC, the cost of spTMS on cSI is more obvious in the cross-modal task than that in the unimodal task (Fig. 6C). As the cost can be recognized as information processed in a corresponding brain area, the current results suggest that more information is processed in cSI than that in cPPC in the cross-modal task at STP 300 ms, but information processed in those two areas is more or less equal in the unimodal task at 300 ms. The previous human electroencephalogram studies have indicated that two parietal event-related potential components (located at about 300 and 600 ms, respectively, after the onset of tactile stimuli) are associated with cross-modal information processing during WM (Ohara et al., 2006, 2008). Here, the causal contribution of SI and PPC to both unimodal and cross-modal WM has been validated by stimulating them at the corresponding time points. Furthermore, through comparisons in processing patterns of sensory information in the delay period between unimodal and cross-modal tasks, the authors believe that cross-modal information transfer does occur in the early delay, most likely during the period between 300 and 600 ms (Ohara et al., 2006, 2008).
Author contributions
Conclusions In the present study, the roles of SI and PPC during tactile unimodal WM were investigated. spTMS applied over cSI and cPPC significantly decreased the accuracy of task performance, compared with the control stimulation on iSI. However, stimulation at all time
Y.K., Y.-D.Z. conceived the study; D.Z. collected the data; Y.K. and D.Z. analysed the data; Y.K., D.Z., M.B. and Y.-D.Z. wrote the manuscript.
References Baddeley, A. (2003) Working memory: looking back and looking forward. Nat. Rev. Neurosci., 4, 829–839. Ballard, D.H. (1986) Cortical connections and parallel processing: structure and function. Behav. Brain Sci., 9, 67–90. Benjamini, Y. & Yekutieli, D. (2001) The control of the false discovery rate in multiple testing under dependency. Ann. Stat., 29, 1165–1188. Bigelow, J., Rossi, B. & Poremba, A. (2014) Neural correlates of short-term memory in primate auditory cortex. Front. Neurosci., 8, 250. Curtis, C.E. & Lee, D. (2010) Beyond working memory: the role of persistent activity in decision making. Trends Cogn. Sci., 14, 216–222. D’Ardenne, K., Eshel, N., Luka, J., Lenartowicz, A., Nystrom, L.E. & Cohen, J.D. (2012) Role of prefrontal cortex and the midbrain dopamine system in working memory updating. Proc. Natl. Acad. Sci. USA, 109, 19900–19909. Davis, T. & Poldrack, R.A. (2013) Measuring neural representations with fMRI: practices and pitfalls. Ann. NY Acad. Sci., 1296, 108–134. Ester, E.F., Anderson, D.E., Serences, J.T. & Awh, E. (2013) A neural measure of precision in visual working memory. J. Cognitive Neurosci., 25, 754–761. Fuster, J.M. (2001) The prefrontal cortex–an update: time is of the essence. Neuron, 30, 319–333. Fuster, J.M. & Alexander, G.E. (1971) Neuron activity related to short-term memory. Science, 173, 652–654. Fuster, J.M. & Bressler, S.L. (2012) Cognit activation: a mechanism enabling temporal integration in working memory. Trends Cogn. Sci., 16, 207–218. Goff, G.D. (1967) Differential discrimination of frequency of cutaneous mechanical vibration. J. Exp. Psychol., 74, 294–299. Goldman-Rakic, P.S. (1988) Topography of cognition: parallel distributed networks in primate association cortex. Annu. Rev. Neurosci., 11, 137– 156.
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
SI and PPC in tactile WM processing 1911 Goldman-Rakic, P.S. (1995) Cellular basis of working memory. Neuron, 14, 477–485. Hannula, H., Neuvonen, T., Savolainen, P., Hiltunen, J., Ma, Y.-Y., Antila, H., Salonen, O., Carlson, S. & Pertovaara, A. (2010) Increasing top-down suppression from prefrontal cortex facilitates tactile working memory. NeuroImage, 49, 1091–1098. Harris, J.A., Miniussi, C., Harris, I.M. & Diamond, M.E. (2002) Transient storage of a tactile memory trace in primary somatosensory cortex. J. Neurosci., 22, 8720–8725. Harrison, S.A. & Tong, F. (2009) Decoding reveals the contents of visual working memory in early visual areas. Nature, 458, 632–635. Haynes, J.-D. & Rees, G. (2006) Decoding mental states from brain activity in humans. Nat. Rev. Neurosci., 7, 523–534. Inui, K., Wang, X., Tamura, Y., Kaneoke, Y. & Kakigi, R. (2004) Serial processing in the human somatosensory system. Cereb. Cortex, 14, 851–857. Iwamura, Y. (1998) Hierarchical somatosensory processing. Curr. Opin. Neurobiol., 8, 522–528. Kaas, A.L., van Mier, H., Visser, M. & Goebel, R. (2013) The neural substrate for working memory of tactile surface texture. Hum. Brain Mapp., 34, 1148–1162. Kalberlah, C., Villringer, A. & Pleger, B. (2013) Dynamic causal modeling suggests serial processing of tactile vibratory stimuli in the human somatosensory cortex–an fMRI study. NeuroImage, 74, 164–171. Kammer, T., Beck, S., Thielscher, A., Laubis-Herrmann, U. & Topka, H. (2001) Motor thresholds in humans: a transcranial magnetic stimulation study comparing different pulse waveforms, current directions and stimulator types. Clin. Neurophysiol., 112, 250–258. Karhu, J. & Tesche, C.D. (1999) Simultaneous early processing of sensory input in human primary (SI) and secondary (SII) somatosensory cortices. J. Neurophysiol., 81, 2017–2025. Klingberg, T., Kawashima, R. & Roland, P.E. (1996) Activation of multimodal cortical areas underlies short-term memory. Eur. J. Neurosci., 8, 1965–1971. Koch, K.W. & Fuster, J.M. (1989) Unit activity in monkey parietal cortex related to haptic perception and temporary memory. Exp. Brain Res., 76, 292–306. Ku, Y., Zhao, D., Hao, N., Hu, Y., Bodner, M. & Zhou, Y.-D. (2015) Sequential roles of primary somatosensory cortex and posterior parietal cortex in tactile-visual cross-modal working memory: a single-pulse Transcranial Magnetic Stimulation (spTMS) Study. Brain Stimul., 8, 88–91. Liang, M., Mouraux, A. & Iannetti, G.D. (2011) Parallel processing of nociceptive and non-nociceptive somatosensory information in the human primary and secondary somatosensory cortices: evidence from dynamic causal modeling of functional magnetic resonance imaging data. J. Neurosci., 31, 8976–8985. Meehan, S.K., Linsdell, M.A., Handy, T.C. & Boyd, L.A. (2011) Interhemispheric enhancement of somatosensory cortical excitability through contralateral repetitive transcranial magnetic stimulation. Clin. Neurophysiol., 122, 1637–1644. Miller, E.K. & Cohen, J.D. (2001) An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci., 24, 167–202. Mull, B.R. & Seyal, M. (2001) Transcranial magnetic stimulation of left prefrontal cortex impairs working memory. Clin. Neurophysiol., 112, 1672– 1675. M€uri, R.M., Vermersch, A.I., Rivaud, S., Gaymard, B. & Pierrot-Deseilligny, C. (1996) Effects of single-pulse transcranial magnetic stimulation over the
prefrontal and posterior parietal cortices during memory-guided saccades in humans. J. Neurophysiol., 76, 2102–2106. Ohara, S., Lenz, F.A. & Zhou, Y.-D. (2006) Modulation of somatosensory event-related potential components in a tactile-visual cross-modal task. Neuroscience, 138, 1387–1395. Ohara, S., Wang, L., Ku, Y., Lenz, F.A., Hsiao, S.S., Hong, B. & Zhou, Y.D. (2008) Neural activities of tactile cross-modal working memory in humans: an event-related potential study. Neuroscience, 152, 692–702. Oliveri, M., Turriziani, P., Carlesimo, G.A., Koch, G., Tomaiuolo, F., Panella, M. & Caltagirone, C. (2001) Parieto-frontal interactions in visual-object and visual-spatial working memory: evidence from transcranial magnetic stimulation. Cereb. Cortex, 11, 606–618. Pasalar, S., Ro, T. & Beauchamp, M.S. (2010) TMS of posterior parietal cortex disrupts visual tactile multisensory integration. Eur. J. Neurosci., 31, 1783–1790. Pascual-Leone, A., Walsh, V. & Rothwell, J. (2000) Transcranial magnetic stimulation in cognitive neuroscience–virtual lesion, chronometry, and functional connectivity. Curr. Opin. Neurobiol., 10, 232–237. Pasternak, T. & Greenlee, M.W. (2005) Working memory in primate sensory systems. Nat. Rev. Neurosci., 6, 97–107. Ploner, M., Schmitz, F., Freund, H.J. & Schnitzler, A. (1999) Parallel activation of primary and secondary somatosensory cortices in human pain processing. J. Neurophysiol., 81, 3100–3104. Pons, T.P., Garraghty, P.E., Friedman, D.P. & Mishkin, M. (1987) Physiological evidence for serial processing in somatosensory cortex. Science, 237, 417–420. Pons, T.P., Garraghty, P.E. & Mishkin, M. (1992) Serial and parallel processing of tactual information in somatosensory cortex of rhesus monkeys. J. Neurophysiol., 68, 518–527. Preuschhof, C., Heekeren, H.R., Taskin, B., Schubert, T. & Villringer, A. (2006) Neural correlates of vibrotactile working memory in the human brain. J. Neurosci., 26, 13231–13239. Romo, R. & Salinas, E. (2003) Flutter discrimination: neural codes, perception, memory and decision making. Nat. Rev. Neurosci., 4, 203–218. Salinas, E., Hernandez, A., Zainos, A. & Romo, R. (2000) Periodicity and firing rate as candidate neural codes for the frequency of vibrotactile stimuli. J. Neurosci., 20, 5503–5515. Savini, N., Brunetti, M., Babiloni, C. & Ferretti, A. (2012) Working memory of somatosensory stimuli: an fMRI study. Int. J. Psychophysiol., 86, 220– 228. Serences, J.T., Ester, E.F., Vogel, E.K. & Awh, E. (2009) Stimulus-specific delay activity in human primary visual cortex. Psychol. Sci., 20, 207–214. Super, H., Spekreijse, H. & Lamme, V.A. (2001) A neural correlate of working memory in the monkey primary visual cortex. Science, 293, 120– 124. Zhang, H.Q., Murray, G.M., Turman, A.B., Mackie, P.D., Coleman, G.T. & Rowe, M.J. (1996) Parallel processing in cerebral cortex of the marmoset monkey: effect of reversible SI inactivation on tactile responses in SII. J. Neurophysiol., 76, 3633–3655. Zhou, Y.-D. & Fuster, J.M. (1996) Mnemonic neuronal activity in somatosensory cortex. Proc. Natl. Acad. Sci. USA, 93, 10533–10537. Zhou, Y.-D. & Fuster, J.M. (2000) Visuo-tactile cross-modal associations in cortical somatosensory cells. Proc. Natl. Acad. Sci. USA, 97, 9777–9782. Zhou, Y.-D., Ardestani, A. & Fuster, J.M. (2007) Distributed and associative working memory. Cereb. Cortex, 17(Suppl. 1), i77–i87.
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
doi:10.1111/ejn.12950
COGNITIVE NEUROSCIENCE
Cooperative processing in primary somatosensory cortex and posterior parietal cortex during tactile working memory Yixuan Ku,1,2,3,* Di Zhao,1,* Mark Bodner4 and Yong-Di Zhou3,5 1
The Key Lab of Brain Functional Genomics, MOE & STCSM, Institute of Cognitive Neuroscience, 3663, North Zhongshan Road, School of Psychology and Cognitive Science, East China Normal University, Shanghai 200062, China 2 Departments of Neurology, Physiology and Psychiatry, University of California, San Francisco, CA, USA 3 NYU-ECNU Institute of Brain and Cognitive Science, NYU Shanghai and Collaborative Innovation Center for Brain Science, Shanghai, China 4 MIND Research Institute, Irvine, CA, USA 5 Krieger Mind/Brain Institute, Johns Hopkins University, 3400 N. Charles Street, 338 Krieger Hall, Baltimore, MA 21218, USA Keywords: cooperative processing, posterior parietal cortex, primary somatosensory cortex, single-pulse transcranial magnetic stimulation, tactile working memory
Abstract In the present study, causal roles of both the primary somatosensory cortex (SI) and the posterior parietal cortex (PPC) were investigated in a tactile unimodal working memory (WM) task. Individual magnetic resonance imaging-based single-pulse transcranial magnetic stimulation (spTMS) was applied, respectively, to the left SI (ipsilateral to tactile stimuli), right SI (contralateral to tactile stimuli) and right PPC (contralateral to tactile stimuli), while human participants were performing a tactile-tactile unimodal delayed matching-to-sample task. The time points of spTMS were 300, 600 and 900 ms after the onset of the tactile sample stimulus (duration: 200 ms). Compared with ipsilateral SI, application of spTMS over either contralateral SI or contralateral PPC at those time points significantly impaired the accuracy of task performance. Meanwhile, the deterioration in accuracy did not vary with the stimulating time points. Together, these results indicate that the tactile information is processed cooperatively by SI and PPC in the same hemisphere, starting from the early delay of the tactile unimodal WM task. This pattern of processing of tactile information is different from the pattern in tactile-visual cross-modal WM. In a tactile-visual cross-modal WM task, SI and PPC contribute to the processing sequentially, suggesting a process of sensory information transfer during the early delay between modalities.
Introduction Working memory (WM) refers to cognitive processes that manipulate and maintain sensory information for a short term to guide goal-directed behavioral actions (Baddeley, 2003). Persistent activation of neurons in association cortices [e.g. the prefrontal cortex (PFC) and the posterior parietal cortex (PPC)] and sensory cortices during a delay period of a WM task has been indicated by studies to represent the neural mechanisms underlying those maintaining and manipulating processes (see reviews in Pasternak & Greenlee, 2005; Curtis & Lee, 2010). However, the temporal dependence of these areas in WM processes still remains unclear. Hierarchical (Fuster, 2001) and parallel (Ballard, 1986; GoldmanRakic, 1988) models have been proposed for WM information
Correspondence: Dr Yixuan Ku, 1The Key Lab of Brain Functional Genomics, as above. E-mail: [email protected] Dr Yong-Di Zhou, 5Krieger Mind/Brain Institute, as above. E-mail: [email protected] *Y.K. and D.Z. contributed equally to this work. Received 10 February 2015, revised 25 April 2015, accepted 13 May 2015
processing, and both gain a plethora of supports (Fuster & Bressler, 2012). Nevertheless, in the somatosensory domain, the PPC is more frequently recognised to be a higher stage compared with the primary somatosensory cortex (SI; Iwamura, 1998; Fuster, 2001; Fuster & Bressler, 2012). However, to date there is still a lack of evidence for the hierarchical relationship between PPC and SI in tactile WM. Posterior parietal cortex has been reported to play an important role in tactile WM. Earlier neurophysiological work with primates has revealed that neurons in PPC show elevated and selective activities towards tactile stimuli through the delay period (Koch & Fuster, 1989). Recent neuroimaging studies with both positron emission tomography and functional magnetic resonance imaging (MRI) have shown that PPC is activated during WM (Klingberg et al., 1996; Preuschhof et al., 2006; Kaas et al., 2013). Meanwhile, recent neurophysiological data have suggested important roles of lower-level sensory cortices [SI and secondary somatosensory cortex (SII)] in tactile WM (Zhou & Fuster, 1996; Salinas et al., 2000; Romo & Salinas, 2003). The causal role of SI in maintaining vibrotactile memory traces has been reported in a human transcranial magnetic stimulation (TMS) study (Harris et al., 2002). Although the role
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd
1906 Y. Ku et al. played by SI in tactile WM has been welcomed by many researchers, there are still studies that do not support the finding (Hannula et al., 2010). In this present study, PPC was further recruited as a stimulating site, and the roles of SI and PPC were jointly assessed in a tactile unimodal delayed matching-to-sample task. Single-pulse TMS (spTMS) was used to stimulate SI and PPC at distinct temporal stages in the task. This stimulation is non-invasive and widely used to test the casual influence of brain activity over behavioral performance, especially regarding the time course of the involvement of a specific cortical area in a given cognitive process (M€ uri et al., 1996; Pascual-Leone et al., 2000). By using spTMS, sequential roles of SI and PPC in a tactilevisual cross-modal WM task (Ku et al., 2015) have recently been shown. Here, whether such a sequential relationship between SI and PPC would also exist in tactile unimodal WM was tested.
Materials and methods Participants Fifteen individuals participated in the experiment (11 females and four males, mean age = 22.7 years). None of the participants had neurological, psychiatric or other relevant medical problems. They did not have any contraindication to TMS. They were given written consent forms detailing the potential risks of the TMS experiment and signed to participate. After the experiment, they were paid for compensation. Twelve participants (out of 15) have participated in the cross-modal experiment in the previous study (Ku et al., 2015). All experimental protocols were approved by the Institutional Review Board of East China Normal University. Experimental paradigm A participant sat in a comfortable chair in front of a 17-inch CRT monitor (IBM C220P CRT; resolution ratio = 800 9 600 pixels; monitor refresh interval = 16.66 ms), with a chin support situated 0.87 m away from the monitor. The experiment consisted of a tactile-tactile unimodal delayed matching-to-sample task. In this task, a complete trial included a series of events as illustrated in Fig. 1, starting with a gray fixation-cross (lasting 1000–1500 ms) at the center of the monitor. The participant was asked to focus his/her attention on the gray cross. When the gray cross turned red, a tactile vibration (S-1, 200 ms duration) was applied to the left index finger of the participant. Frequencies of the vibration were 30, 40, 60 or
100 Hz, based on the equal sensation contours for vibration (Goff, 1967). A delay interval of 1000 ms followed S-1 when the central cross turned back to gray (offset of S-1). During the delay, the participant was instructed to focus his/her attention on the central cross again, and maintain in his/her mind the vibration frequency of S-1. The delay ended with the onset of a second stimulus (S-2, 200 ms duration) that was also a tactile vibration from the same vibration sets of S-1. The cross at the center of the monitor screen turned red during the presence of S-2. The participant reported whether S-2 matched S-1 by pressing one of two buttons as quickly and accurately as possible. The button-assignment (e.g. the right index finger for match, the right middle finger for non-match) was counterbalanced across participants. Both the participant’s choice and response time (RT; the span from the offset of S-2 to the choice) were recorded. A complete trial ended with the choice. There was an interval between two trials, which was randomly set between 3 and 4 s. Each participant performed three training blocks (with feedback of performance information, a correct or an incorrect response) to learn the task, and subsequently performed six experimental TMS blocks (without feedback). Each of these blocks consisted of 48 trials (24 ‘match’ trials and 24 ‘non-match’ trials, randomly arranged). Tactile stimulator A permanent magnetic vibrator (LDS V101 vibrator; probe diameter, 6.4 mm) driven by a LDS PA25E Power Amplifier (Br€ uel & Kjær Sound & Vibration. Measurement A/S, Denmark) was used as a tactile stimulator. The amplitude of a vibration at different frequencies was restricted to the same level (vertical displacement, 0.049 inches). The participant rested his/her left index finger at a fixed location on the vibrator. TMS protocol A similar TMS protocol was operated as in the previous study (Ku et al., 2015). TMS was delivered with a Magstim Rapid2 stimulator and a 70-mm figure-eight-shaped coil (The Magstim Company, Whitland, UK). High-resolution anatomical T1-weighted MRIs were acquired with a 3T Siemens Trio MRI scanner for each participant (TR = 2530 ms, TE = 2.340 ms, flip angle = 35°, FOV = 256 mm, 170 slices, 1 mm thickness) at the ECNU MRI Research Center. These images were then imported into the BrainSightTM neuronavigation software (BrainSight 2.0, Rogue Research, Montreal, Canada)
Fig. 1. Experimental paradigm. In this tactile-tactile unimodal delayed matching-to-sample task, the participant is asked to report whether the second stimulus matches the first stimulus in each trial. Single-pulse transcranial magnetic stimulation (spTMS) is delivered during the delay period at 300/600/900 ms after the onset of the first stimulus. © 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
SI and PPC in tactile WM processing 1907 to allow for stereotaxic registration of the TMS coil with the participant’s brain areas. The resting motor threshold that represented the cortical excitability threshold for each participant was measured by the minimal intensity necessary to elicit a visible movement of the relaxed abductor pollicis brevis muscle of the right hand in five out of 10 TMS pulses applied over the motor hand area in the left hemisphere (Kammer et al., 2001). The strength of TMS for each participant was then calculated as 110% of the motor threshold (mean: 67.6% maximum machine output; standard deviation: 8.2%). The motor hand area was also marked for each participant within the Brainsight software. The left SI (ipsilateral SI to the tactile stimuli, iSI) was localised as a point 2 cm posterior to the motor spot (Meehan et al., 2011). The mean Talairach coordinates of iSI across participants were as follows: x = 34 mm, y = 36 mm, z = 51 mm (see individual locations in Fig. 2, yellow balls). The contralateral SI (cSI) was localized as a mirror point of iSI towards the sagittal plane of the brain. The mean Talairach coordinates of cSI were: x = 34 mm, y = 38 mm,
z = 51 mm (see individual locations in Fig. 2, red balls). The contralateral PPC (cPPC) was localized as the spot about 1 cm posterior and 2 cm lateral from cSI (Pasalar et al., 2010). The mean Talairach coordinates of cPPC were: x = 45 mm, y = 46 mm, z = 50 mm (see individual locations in Fig. 2, blue balls), which was located in the right inferior parietal lobule. The TMS coil was held to the head of the participant with a custom coil holder, and was orientated about 45° clockwise away from the sagittal line at all stimulating sites and tangentially to the scalp. spTMS was then delivered, respectively, over the iSI, cSI and cPPC (two experimental blocks for each stimulating brain site) when the participant performed the task (the order of blocks assigned randomly for each participant). Three different stimulating time points (STPs: 300, 600 and 900 ms after the onset of S-1) for spTMS were the same as in the previous study (Ku et al., 2015). Within each block, 48 trials were equally divided into three groups in line with the STPs and, in each group, the four vibration frequencies were arranged pseudo-randomly, with each of them having four trials. All 48 trials in a single block were shuffled randomly for each participant. Paradigms and TMS were programmed using Matlab Psychtoolbox-3 (www.psychtoolbox.org) based on Matlab R2010b (MathWorks, MA, USA). White noise To attenuate influences from the noise that accompanied vibrations and TMS, earplugs were used and, in addition, white noise (80 dB) was generated using Adobe Audition 3.0 (Adobe Systems, USA) and delivered during the whole experiment through two loudspeakers placed on two sides of the CRT, respectively. Statistical analysis
Fig. 2. The stimulating sites across participants [yellow balls, ipsilateral primary somatosensory cortex (iSI); red balls, contralateral (cSI); blue balls, contralateral posterior parietal cortex (cPPC)].
A
The accuracy of task performance was defined as the percentage of correct responses in each condition, and the RT was defined as the interval between the offset of S-2 and the subject’s correct response. They were analysed by a two-way repeated-measures analysis of variance (RM-ANOVA), with LOC (stimulating sites: iSI, cSI and cPPC) and STP as within-subject factors. A two-sided paired Student’s t-test was performed whenever there were significant main effects or interactions. Multiple comparisons were corrected using the false discovery rate (FDR) method (Benjamini & Yekutieli, 2001).
B
Fig. 3. (A) Accuracy and (B) response time (RT) of behavioral performance across conditions. Dashed lines represent the mean of baseline performance in the last training block. The gray shaded area indicates the standard error of baseline performance. Error bars denote SEs. © 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
1908 Y. Ku et al. The participant’s performance in the third training block was extracted as a baseline without stimulation. TMS effects on performance were also compared between situations of TMS stimulation and baseline by using the paired t-test. Data from the 12 participants who participated in both unimodal and cross-modal studies were used to compare their performance between the studies. The baseline accuracy of performance in each study was used to normalise the accuracy data. The spTMS cost was defined as the percentage of drop-off from the baseline accuracy to the accuracy when spTMS was applied (as illustrated in the following equation).
CostAccuracy ¼
Accuracybaseline AccuracyspTMS 100% Accuracybaseline
All statistical analyses were performed with Statistica 6.0 (StatSoft, Tulsa, OK, USA).
Results After three blocks of training, participants performed the tasks with accuracy above 80% in the last training block (Fig. 3A). The task performance under all conditions was illustrated in Fig. 3 (accuracy, Fig. 3A; RT, Fig. 3B). RM-ANOVA for accuracy revealed a significant main effect of LOC (F2,28 = 9.25, P = 0.0008, gp2 = 0.398), but neither a significant main effect of STP (F2,28 = 0.529, P = 0.595, gp2 = 0.036) nor an interaction between LOC and STP (F4,56 = 0.233, P = 0.919, gp2 = 0.016) was observed. Post hoc analysis for LOC revealed that compared with iSI, stimulation on cSI or cPPC significantly (both FDR corrected, P < 0.05) decreased the accuracy of performance (Fig. 4). Further comparisons against baseline showed that stimulation on cSI or cPPC also significantly (both FDR corrected, P < 0.01) deteriorated the accuracy (Fig. 4), but stimulation on iSI did not (t14 = 1.39, P = 0.186).
RM-ANOVA for RT revealed neither a significant main effect of the factors, nor an interaction between them (Fig. 3B). To compare the results in the present unimodal study with those in the previous cross-modal study (Ku et al., 2015), first it was assessed whether there was any significant difference in baseline performance between the unimodal and cross-modal tasks. The twosided paired t-test revealed that there was no significant difference in baseline accuracy between those two tasks (t11 = 1.23, P = 0.24), indicating that the task difficulty was similar between them. However, baseline RT in the cross-modal task (mean SEM: 661 51 ms) was significantly longer than in the unimodal task (509 45 ms; t11 = 2.86, P = 0.01), which was likely due to a distinct sensory modality that S-2 (visual stimuli) had in the crossmodal task. Therefore, this study mainly focused on accuracy in conducting a comparison between unimodal and cross-modal task performance. spTMS showed sequentially deteriorative effects on cSI at STP 300 ms and on cPPC at 600 ms in the cross-modal task (Ku et al., 2015), but in the unimodal task (present study), spTMS showed deteriorative effects on cSI and cPPC across all STPs (Fig. 5). Thus, the cost of spTMS in these two conditions was assessed with a three-way RM-ANOVA, using TASK (cross-modal and unimodal), LOC (cSI and cPPC) and STP (300 and 600 ms) as within-subject factors. A marginal TASK*LOC*STP (F1,11 = 4.45, P = 0.0587, gp2 = 0.288) interaction was spotted. Further two-way RM-ANOVA with TASK and LOC revealed that, at STP 300 ms, there was a significant main effect of LOC (F1,11 = 8.52, P = 0.014, gp2 = 0.437) and a significant interaction between TASK and LOC (F1,11 = 14.4, P = 0.003, gp2 = 0.567), while at STP 600 ms, neither significant main effect nor interaction was observed. The individual spTMS cost at STP 300 ms was shown in Fig. 6A and B. Costs were higher under cSI stimulation than under cPPC stimulation in the cross-modal task (Fig. 6A), while they varied in the unimodal task (Fig. 6B). Cost differences between cSI and cPPC (cSI – cPPC) of all participants were compared between the two tasks, a significantly higher average cost was observed in the crossmodal task (t11 = 3.797, P = 0.003, as in Fig. 6C).
Discussion The main finding of the present study is that SI and PPC in the same hemisphere are both essential in tactile information processing in the delay period of a tactile unimodal WM task. Those two cortical areas work together in the processing, starting from the early delay (not later than 300 ms after the onset of tactile sample stimuli) through a period of at least 600 ms. Further, based on detailed comparisons of the current results of TMS effects received from the tactile unimodal WM task with the previous findings of TMS effects received from a tactile-visual cross-modal WM task (Ku et al., 2015), the authors believe that it is a reasonable assumption that the cross-modal information transfer occurs during the period between 300 and 600 ms after the onset of the sample stimulus (a duration of 200 ms). Involvement of SI and PPC during tactile WM
Fig. 4. The main effect of stimulating location. Dashed lines represent the mean of baseline performance in the last training block. The gray shaded area indicates the standard error of baseline performance. [false discovery rate (FDR) corrected, *P < 0.05, **P < 0.01; error bars denote SEs].
In most experiments of spTMS on frontal and parietal areas to date, the stimulation disturbed brain activities and downgraded performance in WM tasks (M€ uri et al., 1996; Mull & Seyal, 2001; Oliveri et al., 2001; Harris et al., 2002; D’Ardenne et al., 2012). Consistent with those studies, manipulation of TMS in the current study shows significant impairment of participants’ performance at all STPs,
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
SI and PPC in tactile WM processing 1909 A
B
Fig. 5. Comparisons of stimulating costs between cross-modal and unimodal tasks. Accuracy cost of (A) contralateral posterior parietal cortex (cPPC) vs. (B) contralateral primary somatosensory cortex (cSI).
A
B
C
Fig. 6. Comparisons of the stimulating cost between cross-modal and unimodal tasks at stimulating time point (STP) 300 ms. Scatter plot across participants in (A) the cross-modal task and (B) the unimodal task [accuracy cost of contralateral posterior parietal cortex (cPPC) vs. contralateral primary somatosensory cortex (cSI)]. (C) Differences in cost between cSI and cPPC across participants in cross-modal and unimodal tasks. Each circle represents an individual participant from whom data were received from both tasks in the study [data from the additional three participants for the unimodal study are only shown in (B) as squares].
indicating that both cSI and cPPC are task-related brain areas and play causal roles for tactile WM processing. The results are reminiscent of those neuroimaging findings on the involvement of PPC during tactile WM maintenance (Klingberg et al., 1996; Preuschhof et al., 2006; Savini et al., 2012; Kaas et al., 2013). The effects on performance are not likely due to clicking sounds induced by TMS because, on all stimulating sites, noise produced by stimulation is essentially at the same level. In a previous empirical study, Harris et al. (2002) used spTMS to explore the causal role of cSI in a tactile WM task. spTMS on cSI exhibited more severe decline of performance than on iSI, when stimulation was applied at 300 ms and 600 ms after the offset of tactile stimuli, but not at 900 and 1200 ms (Harris et al., 2002). The current results of cSI stimulation coincide with their findings at STPs 300 and 600 ms. However, a significant effect with spTMS on cSI at STP 900 ms has also been observed. This result may not contradict with Harris and colleagues’ findings, as the duration of tactile stimuli used in the current study (200 ms) is much shorter than theirs (1000 ms). In fact, the STPs in the current study all lie within the duration of their stimuli. Cooperative processing in SI and PPC during unimodal WM It has long been debated for somatosensory information processing at earlier sensory stages, whether serial or parallel processing dominates within SI and SII (Pons et al., 1987, 1992; Zhang et al., 1996; Iwamura, 1998; Karhu & Tesche, 1999; Ploner et al., 1999;
Inui et al., 2004; Liang et al., 2011; Kalberlah et al., 2013). Meanwhile, PPC has been widely accepted as a later stage of somatosensory processing and WM (Iwamura, 1998; Fuster, 2001; Fuster & Bressler, 2012). In the present study, a non-significant main effect of STP in both cSI and cPPC has been observed, suggesting that SI and PPC remain cooperative in processing sensory information across the delay period (at least starting at 300 ms after the onset of the sensory cue). Compared with the sequential processing pattern in the previous cross-modal WM study (Ku et al., 2015), the processing pattern here in the unimodal WM indicates distinct neural mechanisms. The unimodal WM does not demand an information transfer from one sensory modality to another, but requires maintenance of information in the same modality across the delay period and, therefore, a lag of 300 ms is not observed in TMS effects on PPC in the unimodal task. Apparently, this 300 ms delay in the previous cross-modal study is very likely for the information transfer between modalities. Further, stimulation of either cSI or cPPC shows TMS effects on task performance at all STPs, which indicates that to properly keep tactile information during the delay period, both SI and PPC are required. That is, sensory information is represented in both areas, each of which is critical in WM. These results further support a distributed WM network proposal (Zhou et al., 2007; Fuster & Bressler, 2012). Future study to explore the temporal profile of the earlier stage of sensory processing (during the presentation of stimuli) will be of interest, and future research with an extension of the delay period to
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
1910 Y. Ku et al. test if the conclusion of cooperative processing in SI and PPC during a long delay still holds will also be interesting. Primary sensory cortices in WM maintenance Compared with studies of sustained, elevated delay activity in the PFC (Fuster & Alexander, 1971; Goldman-Rakic, 1995; Miller & Cohen, 2001; Curtis & Lee, 2010), fewer studies have revealed sustained delay activity in primary sensory cortices in WM tasks (Zhou & Fuster, 1996, 2000; Super et al., 2001; Romo & Salinas, 2003; Bigelow et al., 2014). However, recent advances in neuroimaging have presented the possibility to accurately decode the cortical representation of sensory information during the WM (see reviews in Haynes & Rees, 2006; Davis & Poldrack, 2013), and thus revealed with multivariate pattern analysis content-specific representations in primary visual regions during WM (Harrison & Tong, 2009; Serences et al., 2009). Harrison and Tong have shown that even if the overall delay activity is low in human visual cortices, orientations held in WM can still be clearly decoded from activity patterns (Harrison & Tong, 2009). A recent study by Ester et al. (2013) has further revealed that information decoded in the primary sensory cortex is restricted for those remembered features. All those results indicate that primary sensory cortices are also involved in maintenance of sensory information during WM (Pasternak & Greenlee, 2005), and the current results provide further evidence for such involvement.
points did not reveal different effects between cSI and cPPC. Together, these results indicate cooperative processing in SI and PPC starting at an early stage during tactile unimodal WM. Comparisons in the TMS effects between unimodal and cross-modal tasks lead to a reasonable assumption that the cross-modal information transfer likely occurs during the period between 300 and 600 ms.
Acknowledgements The authors thank Hongyuan Li for his help on the interface of the tactile vibrator, and Liping Wang for his comments on the earlier proposals. This work was supported by National Key Fundamental Research (973) Program (2013CB329501), National Natural Science Foundation of China (31100742), Shanghai Committee of Science and Technology (15ZR1410600), and a research fund from the M.I.N.D. Research Institute, California.
Abbreviations cPPC, contralateral posterior parietal cortex; cSI, contralateral primary somatosensory cortex; FDR, false discovery rate; iSI, ipsilateral primary somatosensory cortex; MRI, magnetic resonance imaging; PFC, prefrontal cortex; PPC, posterior parietal cortex; RT, response time; SI, primary somatosensory cortex; SII, secondary somatosensory cortex; spTMS, singlepulse transcranial magnetic stimulation; STP, stimulating time point; TMS, transcranial magnetic stimulation; WM, working memory.
Conflict of interest Primary somatosensory cortices in cross-modal information transfer
None to declare.
It is notable that SI collaborates with PPC in different ways between the unimodal and cross-modal tasks, especially at 300 ms after the onset of the sample stimulus. In the cross-modal task, stimulation of cSI always disrupts performance more severe than that of cPPC (Fig. 6A), while in the unimodal task, stimulation of those two areas shows compatible degradation of performance (Fig. 6B). When contrasted with cPPC, the cost of spTMS on cSI is more obvious in the cross-modal task than that in the unimodal task (Fig. 6C). As the cost can be recognized as information processed in a corresponding brain area, the current results suggest that more information is processed in cSI than that in cPPC in the cross-modal task at STP 300 ms, but information processed in those two areas is more or less equal in the unimodal task at 300 ms. The previous human electroencephalogram studies have indicated that two parietal event-related potential components (located at about 300 and 600 ms, respectively, after the onset of tactile stimuli) are associated with cross-modal information processing during WM (Ohara et al., 2006, 2008). Here, the causal contribution of SI and PPC to both unimodal and cross-modal WM has been validated by stimulating them at the corresponding time points. Furthermore, through comparisons in processing patterns of sensory information in the delay period between unimodal and cross-modal tasks, the authors believe that cross-modal information transfer does occur in the early delay, most likely during the period between 300 and 600 ms (Ohara et al., 2006, 2008).
Author contributions
Conclusions In the present study, the roles of SI and PPC during tactile unimodal WM were investigated. spTMS applied over cSI and cPPC significantly decreased the accuracy of task performance, compared with the control stimulation on iSI. However, stimulation at all time
Y.K., Y.-D.Z. conceived the study; D.Z. collected the data; Y.K. and D.Z. analysed the data; Y.K., D.Z., M.B. and Y.-D.Z. wrote the manuscript.
References Baddeley, A. (2003) Working memory: looking back and looking forward. Nat. Rev. Neurosci., 4, 829–839. Ballard, D.H. (1986) Cortical connections and parallel processing: structure and function. Behav. Brain Sci., 9, 67–90. Benjamini, Y. & Yekutieli, D. (2001) The control of the false discovery rate in multiple testing under dependency. Ann. Stat., 29, 1165–1188. Bigelow, J., Rossi, B. & Poremba, A. (2014) Neural correlates of short-term memory in primate auditory cortex. Front. Neurosci., 8, 250. Curtis, C.E. & Lee, D. (2010) Beyond working memory: the role of persistent activity in decision making. Trends Cogn. Sci., 14, 216–222. D’Ardenne, K., Eshel, N., Luka, J., Lenartowicz, A., Nystrom, L.E. & Cohen, J.D. (2012) Role of prefrontal cortex and the midbrain dopamine system in working memory updating. Proc. Natl. Acad. Sci. USA, 109, 19900–19909. Davis, T. & Poldrack, R.A. (2013) Measuring neural representations with fMRI: practices and pitfalls. Ann. NY Acad. Sci., 1296, 108–134. Ester, E.F., Anderson, D.E., Serences, J.T. & Awh, E. (2013) A neural measure of precision in visual working memory. J. Cognitive Neurosci., 25, 754–761. Fuster, J.M. (2001) The prefrontal cortex–an update: time is of the essence. Neuron, 30, 319–333. Fuster, J.M. & Alexander, G.E. (1971) Neuron activity related to short-term memory. Science, 173, 652–654. Fuster, J.M. & Bressler, S.L. (2012) Cognit activation: a mechanism enabling temporal integration in working memory. Trends Cogn. Sci., 16, 207–218. Goff, G.D. (1967) Differential discrimination of frequency of cutaneous mechanical vibration. J. Exp. Psychol., 74, 294–299. Goldman-Rakic, P.S. (1988) Topography of cognition: parallel distributed networks in primate association cortex. Annu. Rev. Neurosci., 11, 137– 156.
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
SI and PPC in tactile WM processing 1911 Goldman-Rakic, P.S. (1995) Cellular basis of working memory. Neuron, 14, 477–485. Hannula, H., Neuvonen, T., Savolainen, P., Hiltunen, J., Ma, Y.-Y., Antila, H., Salonen, O., Carlson, S. & Pertovaara, A. (2010) Increasing top-down suppression from prefrontal cortex facilitates tactile working memory. NeuroImage, 49, 1091–1098. Harris, J.A., Miniussi, C., Harris, I.M. & Diamond, M.E. (2002) Transient storage of a tactile memory trace in primary somatosensory cortex. J. Neurosci., 22, 8720–8725. Harrison, S.A. & Tong, F. (2009) Decoding reveals the contents of visual working memory in early visual areas. Nature, 458, 632–635. Haynes, J.-D. & Rees, G. (2006) Decoding mental states from brain activity in humans. Nat. Rev. Neurosci., 7, 523–534. Inui, K., Wang, X., Tamura, Y., Kaneoke, Y. & Kakigi, R. (2004) Serial processing in the human somatosensory system. Cereb. Cortex, 14, 851–857. Iwamura, Y. (1998) Hierarchical somatosensory processing. Curr. Opin. Neurobiol., 8, 522–528. Kaas, A.L., van Mier, H., Visser, M. & Goebel, R. (2013) The neural substrate for working memory of tactile surface texture. Hum. Brain Mapp., 34, 1148–1162. Kalberlah, C., Villringer, A. & Pleger, B. (2013) Dynamic causal modeling suggests serial processing of tactile vibratory stimuli in the human somatosensory cortex–an fMRI study. NeuroImage, 74, 164–171. Kammer, T., Beck, S., Thielscher, A., Laubis-Herrmann, U. & Topka, H. (2001) Motor thresholds in humans: a transcranial magnetic stimulation study comparing different pulse waveforms, current directions and stimulator types. Clin. Neurophysiol., 112, 250–258. Karhu, J. & Tesche, C.D. (1999) Simultaneous early processing of sensory input in human primary (SI) and secondary (SII) somatosensory cortices. J. Neurophysiol., 81, 2017–2025. Klingberg, T., Kawashima, R. & Roland, P.E. (1996) Activation of multimodal cortical areas underlies short-term memory. Eur. J. Neurosci., 8, 1965–1971. Koch, K.W. & Fuster, J.M. (1989) Unit activity in monkey parietal cortex related to haptic perception and temporary memory. Exp. Brain Res., 76, 292–306. Ku, Y., Zhao, D., Hao, N., Hu, Y., Bodner, M. & Zhou, Y.-D. (2015) Sequential roles of primary somatosensory cortex and posterior parietal cortex in tactile-visual cross-modal working memory: a single-pulse Transcranial Magnetic Stimulation (spTMS) Study. Brain Stimul., 8, 88–91. Liang, M., Mouraux, A. & Iannetti, G.D. (2011) Parallel processing of nociceptive and non-nociceptive somatosensory information in the human primary and secondary somatosensory cortices: evidence from dynamic causal modeling of functional magnetic resonance imaging data. J. Neurosci., 31, 8976–8985. Meehan, S.K., Linsdell, M.A., Handy, T.C. & Boyd, L.A. (2011) Interhemispheric enhancement of somatosensory cortical excitability through contralateral repetitive transcranial magnetic stimulation. Clin. Neurophysiol., 122, 1637–1644. Miller, E.K. & Cohen, J.D. (2001) An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci., 24, 167–202. Mull, B.R. & Seyal, M. (2001) Transcranial magnetic stimulation of left prefrontal cortex impairs working memory. Clin. Neurophysiol., 112, 1672– 1675. M€uri, R.M., Vermersch, A.I., Rivaud, S., Gaymard, B. & Pierrot-Deseilligny, C. (1996) Effects of single-pulse transcranial magnetic stimulation over the
prefrontal and posterior parietal cortices during memory-guided saccades in humans. J. Neurophysiol., 76, 2102–2106. Ohara, S., Lenz, F.A. & Zhou, Y.-D. (2006) Modulation of somatosensory event-related potential components in a tactile-visual cross-modal task. Neuroscience, 138, 1387–1395. Ohara, S., Wang, L., Ku, Y., Lenz, F.A., Hsiao, S.S., Hong, B. & Zhou, Y.D. (2008) Neural activities of tactile cross-modal working memory in humans: an event-related potential study. Neuroscience, 152, 692–702. Oliveri, M., Turriziani, P., Carlesimo, G.A., Koch, G., Tomaiuolo, F., Panella, M. & Caltagirone, C. (2001) Parieto-frontal interactions in visual-object and visual-spatial working memory: evidence from transcranial magnetic stimulation. Cereb. Cortex, 11, 606–618. Pasalar, S., Ro, T. & Beauchamp, M.S. (2010) TMS of posterior parietal cortex disrupts visual tactile multisensory integration. Eur. J. Neurosci., 31, 1783–1790. Pascual-Leone, A., Walsh, V. & Rothwell, J. (2000) Transcranial magnetic stimulation in cognitive neuroscience–virtual lesion, chronometry, and functional connectivity. Curr. Opin. Neurobiol., 10, 232–237. Pasternak, T. & Greenlee, M.W. (2005) Working memory in primate sensory systems. Nat. Rev. Neurosci., 6, 97–107. Ploner, M., Schmitz, F., Freund, H.J. & Schnitzler, A. (1999) Parallel activation of primary and secondary somatosensory cortices in human pain processing. J. Neurophysiol., 81, 3100–3104. Pons, T.P., Garraghty, P.E., Friedman, D.P. & Mishkin, M. (1987) Physiological evidence for serial processing in somatosensory cortex. Science, 237, 417–420. Pons, T.P., Garraghty, P.E. & Mishkin, M. (1992) Serial and parallel processing of tactual information in somatosensory cortex of rhesus monkeys. J. Neurophysiol., 68, 518–527. Preuschhof, C., Heekeren, H.R., Taskin, B., Schubert, T. & Villringer, A. (2006) Neural correlates of vibrotactile working memory in the human brain. J. Neurosci., 26, 13231–13239. Romo, R. & Salinas, E. (2003) Flutter discrimination: neural codes, perception, memory and decision making. Nat. Rev. Neurosci., 4, 203–218. Salinas, E., Hernandez, A., Zainos, A. & Romo, R. (2000) Periodicity and firing rate as candidate neural codes for the frequency of vibrotactile stimuli. J. Neurosci., 20, 5503–5515. Savini, N., Brunetti, M., Babiloni, C. & Ferretti, A. (2012) Working memory of somatosensory stimuli: an fMRI study. Int. J. Psychophysiol., 86, 220– 228. Serences, J.T., Ester, E.F., Vogel, E.K. & Awh, E. (2009) Stimulus-specific delay activity in human primary visual cortex. Psychol. Sci., 20, 207–214. Super, H., Spekreijse, H. & Lamme, V.A. (2001) A neural correlate of working memory in the monkey primary visual cortex. Science, 293, 120– 124. Zhang, H.Q., Murray, G.M., Turman, A.B., Mackie, P.D., Coleman, G.T. & Rowe, M.J. (1996) Parallel processing in cerebral cortex of the marmoset monkey: effect of reversible SI inactivation on tactile responses in SII. J. Neurophysiol., 76, 3633–3655. Zhou, Y.-D. & Fuster, J.M. (1996) Mnemonic neuronal activity in somatosensory cortex. Proc. Natl. Acad. Sci. USA, 93, 10533–10537. Zhou, Y.-D. & Fuster, J.M. (2000) Visuo-tactile cross-modal associations in cortical somatosensory cells. Proc. Natl. Acad. Sci. USA, 97, 9777–9782. Zhou, Y.-D., Ardestani, A. & Fuster, J.M. (2007) Distributed and associative working memory. Cereb. Cortex, 17(Suppl. 1), i77–i87.
© 2015 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 42, 1905–1911
