Journal of Neurophysiology

Handedness, Dexterity, and Motor Cortical Representations

Jessica A. Bernard, Stephan F. Taylor, Rachael D. Seidler


Motor system organization varies with handedness. However, previous work has focused almost exclusively on direction of handedness (right or left) as opposed to degree of handedness (strength). In the present study, we determined whether measures of interhemispheric interactions and degree of handedness are related to contra- and ipsilateral motor cortical representations. Participants completed a battery of handedness assessments including both handedness preference measures and behavioral measures of intermanual differences in dexterity, a computerized version of the Poffenberger paradigm (PP) to estimate interhemispheric transfer time (IHTT), and they underwent transcranial magnetic stimulation (TMS) mapping of both motor cortices while we recorded muscle activity from the first dorsal interosseous muscle bilaterally. A greater number of ipsilateral motor evoked potentials (iMEPs) were elicited in less lateralized individuals with the number of iMEPs correlated with IHTT. There were no relationships between handedness or lateralization of dexterity and symmetry of contralateral motor representations, although this symmetry was related to IHTT. Finally, IHTT was positively correlated with multiple measures of laterality and handedness. These findings demonstrate that degree of laterality of dexterity is related to the propensity for exhibiting iMEPs and the speed of interhemispheric interactions. However, it is not clear whether iMEPs are directly mediated via ipsilateral corticospinal projections or are transcallosally transmitted.


It is known that handedness is related to the organization of the sensorimotor system. For example, there is some evidence to indicate that left- and right-handers may differ in corpus callosum morphology (Habib et al. 1991; Westerhausen et al. 2004; Witelson 1985) and organization of motor cortical representations (Amunts et al. 1996; Dassonville et al. 1997; Kim et al. 1993; Volkmann et al. 1998). These neuroanatomical differences have been linked to behavior as well with left-handers typically having faster and more accurate interhemispheric transmission than right-handers (Cherbuin and Brinkman 2006a,b; Marzi et al. 1991). However, the majority of work investigating relationships between handedness, neuroanatomy, and behavior has evaluated direction of handedness (left vs. right) as opposed to degree of handedness (how strongly handed one is), although there is evidence to support the importance of the latter (Dassonville et al. 1997; Luders et al. 2010; Siebner et al. 2002; Solodkin et al. 2001).

Organization of motor cortex and handedness

Motor cortex physiology is related to handedness. For example, left- and right-handers show different functional activation patterns in the primary motor cortex (M1) when performing motor tasks. Right-handers show more asymmetrical activation in M1, predominantly in the dominant (left) hemisphere, whereas left-handers show more bilateral activity (Kim et al. 1993; Siebner et al. 2002). Those that are more mixed-handed show the most bilateral M1 activity during the performance of a unimanual task (Dassonville et al. 1997). Additionally, several studies have used transcranial magnetic stimulation (TMS) to investigate the relationship between handedness and motor cortical representations (Cicinelli et al. 1997; Triggs et al. 1999; Wassermann et al. 1992; Wilson et al. 1993). While most studies indicate that there is no difference between the motor cortical representations across the two hemispheres, the majority of studies have been limited to right-handed individuals (Cicinelli et al. 1997; Wassermann et al. 1992; Wilson et al. 1993), although Triggs and colleagues (1999) did find asymmetries in the motor cortex between left- and right-handers. However, these studies did not consider degree of handedness nor did they investigate ipsilateral motor representations.

Interhemispheric communication and handedness

The Poffenberger paradigm (PP) (Marzi 1999; Poffenberger 1912) is thought to rely at least in part on the corpus callosum for task performance (Marzi et al. 1999; Zaidel and Iacaboni 2003). The PP results in a measure of interhemispheric transfer time (IHTT) know as the crossed-uncrossed difference (CUD). CUD is calculated by subtracting uncrossed responses (those where the visual information is processed in the same hemisphere as the motor response) from crossed responses (those where the visual information is processed in the hemisphere opposite to that eliciting the motor response). CUDs are on average ∼3–4 ms and typically range from 1 to 10 ms (Marzi 1999; Marzi et al. 1991), although negative CUD times have also been reported (Marzi et al. 1991; Tettamanti et al. 2002). A meta-analysis of studies using the PP revealed that left-handers had faster IHTT (smaller CUDs) than right-handers (Marzi et al. 1991). Additionally, Cherbuin and Brinkman (2006a,b) observed a relationship between IHTT and handedness. Thus although IHTT is related to direction of handedness, it remains unclear whether there is a relationship across the handedness spectrum.

Corpus callosum and handedness

The corpus callosum is the main conduit for information transfer between the two hemispheres. Although the literature is mixed, there is some evidence that there are differences in callosal size in left- and right-handed individuals. Postmortem data demonstrate that the corpus callosum is larger in left- and mixed-handed individuals, particularly in the midsagittal area (Witelson 1985), which connects the sensorimotor cortices. This is complementary to later work also by Witelson (1989) that indicates that “nonconsistent” right-handers have larger callosal area. In addition, diffusion tensor imaging data indicate that left-handers have higher fractional anisotropy and lower mean diffusivity in the callosum; this may be indicative of greater connectivity between the hemispheres. However, right-handers show larger callosal area (Westerhausen et al. 2004), which is inconsistent with earlier work. Additionally, Habib and colleagues (1991) found that “nonconsistent” [laterality index (LI) less than +80] right-handers had larger callosal area than “consistent” right-handers (LI +80 or higher), suggesting that degree of handedness may also relate to speed and accuracy of interhemispheric interactions. More recently, Luders and colleagues (2010) examined callosal size in relation to degree of handedness. In a large sample, they found that less strongly handed individuals have larger callosa in both the anterior and posterior midbody regions. Thus, though the literature is a bit unclear as to whether or not there are differences in callosal morphometry among handedness groups, at least some of the evidence suggests a neuroanatomical basis for more accurate interhemispheric interactions in left-handers as described in the preceding text.

In sum, it is clear that direction of handedness (left or right) is related to M1 organization, speed, and accuracy of interhemispheric interactions, and structural measures of the corpus callosum. The inter-relationships between these multiple variables remain to be determined, however, particularly in relation to degree of handedness. Although research employing functional neuroimaging has looked at degree of handedness (Dassonville et al. 1997; Lutz et al. 2005), to our knowledge, this has not yet been studied using TMS metrics. Furthermore, it is still an open question as to how interhemispheric communication is related to motor cortical symmetry. In the current study, we evaluate the relationships between motor cortical representations and IHTT across the full range of handedness and laterality of dexterity. Based on prior work looking at functional activity in the primary motor cortex across the handedness spectrum (Dassonville et al. 1997; Kim et al. 1993; Siebner et al. 2002), we predicted that more strongly lateralized individuals (regardless of hand preference) would have more asymmetry between the representations contralateral to each hand. Further, we expected that the representations for less lateralized individuals would exhibit more symmetry across the two hemispheres. We hypothesized that less strongly lateralized individuals would have faster IHTT as well as more ipsilateral motor activity resulting from TMS stimulation during rest and that the two measures would be correlated across participants. Finally, we hypothesized that there would be a relationship between the latency of iMEPs and IHTT, supporting the inter-relatedness of degree of handedness and laterality, interhemispheric interactions, and the existence of ipsilateral motor cortical representations.



Forty-eight participants [age: 20.9 ± 1.9 (SD) yr; 13 male; 18 left handed (self-report)] were recruited from the University of Michigan and the surrounding community and were paid for their participation. Left- and mixed-handed participants were over-recruited so that we would have a broad distribution for our investigation. Fliers specifically targeting left-handed individuals were placed around the University of Michigan campus in addition to the standard recruitment fliers, which did not target any particular hand preference group. Furthermore, we had access to a paid participant database, which included information on hand preference for each individual (left, right, or mixed). We recruited left- and mixed-handed individuals using this list. This study was approved by the University of Michigan IRB, and on enrollment in the study participants signed an IRB-approved informed consent form.

Experimental setup and procedure

Testing occurred on two separate days. On day 1, participants completed a battery of handedness assessments: 1) the Edinburgh Handedness Inventory (13-item score) (Oldfield 1971) to provide a self-report measure of handedness; 2) Tapping Circles and Tapping Squares from the Hand Dominance Test (Steingrüber 1971) to provide a graphomotor measure of dexterity for each hand; 3) the Purdue Pegboard Test (Tiffin and Asher 1948); and 4) a grip strength assessment. The Tapping Circles and Squares tasks required participants to tap dots in either small circles or small squares with a felt tipped pen with both the dominant and nondominant hand (not simultaneously, but on separate trials). The number of circles or squares tapped in 30 s was recorded for each hand and used to calculate a laterality index. The order of hands was counterbalanced across participants. The Purdue Pegboard was administered under several conditions. All required participants to manipulate small pegs using either one or both hands. In the unimanual condition, participants put one peg at a time into a row of small holes. In the bimanual condition, both hands were used simultaneously to put pegs into parallel rows of holes, and in the assembly condition, the hands were used in a more asynchronous fashion to build small spools using pegs, cylinders, and washers. All trials lasted for 30 s and were completed three times. The number of pegs (unimanual condition), the number of pairs of pegs (bimanual condition), and the number of overall pieces (assembly condition) were recorded. Grip strength was measured using a hand dynamometer. Participants were asked to squeeze with maximal force. The performing hand was alternated trial to trial, and three trials were completed for each hand. While the Edinburgh Handedness Inventory assesses preferred hand use, our other measures assess dexterity but can be used to measure upper-limb laterality. Laterality indices (LIs) were calculated for all of these measures as follows: (R − L)/(R + L), where R and L indicate left and right hand responses, respectively. Participants also filled out a health history questionnaire.

Participants performed a computerized version of the PP (Poffenberger 1912), implemented with E-Prime 1.1 software (Psychology Software Tools). Participants sat 55 cm away from the computer screen with their chin resting in a chin rest (Applied Science Laboratories) to control for distance to the screen and gaze location. The stimuli consisted of white dots presented on a black background with a size of 0.61° of visual angle with their center location subtending 6.02° of visual angle to the left or right of a central fixation cross. A fixation cross was presented with a variable foreperiod (500, 650, 800, or 1,000 ms) to prevent anticipatory responses. Stimuli were presented in a random fashion to the left and right visual fields for 50 ms. Participants then had 1,000 ms to make their response before the program advanced to the next trial. Responses were made by pressing the middle button of a response box placed along the body midline with either the right or the left hand. Responses were blocked such that each hand responded for 100 trials before the response hand was switched, with the ordering of left and right hand response blocks counterbalanced across participants. Eight hundred trials were completed (400 responses per hand).

On day 2, a subset of 30 participants from the behavioral session (21 ± 2 yr; 7 male; 14 left handed) underwent a TMS motor-mapping procedure. A larger sample was tested for the behavioral session (n = 48) to provide additional power and because some participants were ineligible to undergo TMS based on screening for potential contraindications. Participants sat comfortably in a chair with their head resting in a chin rest and their hands relaxed. A tight-fitting lycra swim cap was placed on the head to allow for the marking of stimulation locations. Motor-evoked potentials (MEPs) were recorded from the first dorsal interosseous muscle (FDI) of both hands using 4 mm Ag/AgCl electrodes placed on the muscle in a belly-tendon arrangement. A ground electrode was placed on the right wrist. MEP data were recorded and digitized at 2,000 Hz using Biopac hardware and AcqKnowledge software (BIOPAC Systems, Goleta, CA). Our system was carefully shielded by twisting the electrode wires, and all leads were secured in place so that they were not in contact with one another. A Magstim 70 mm figure-eight coil with a Magstim Rapid stimulator (Magstim, Wales, UK) was used for TMS stimulation. The motor hot spot for the FDI muscle for each hand was localized by stimulating at a supra-threshold level of intensity. Resting motor threshold was then determined to the nearest two percent of stimulator output that elicited MEPs of ≥50 μV on three of six consecutive stimulations while the target muscle was at rest (cf. Triggs et al. 1994, 1999). A 6 × 6 grid of points 12 mm apart was used for motor mapping in each hemisphere and was designed to encompass the majority of the motor cortical hand representation. Grid location was determined based on anatomical landmarks such that the top of the grid was placed 2 cm inferolaterally from point Cz (Sparing et al. 2008). The grid extended 4.8 cm in the anterior direction and 1.2 cm in the posterior direction (Fig. 1). Each site was stimulated six times at 110% of motor threshold with ≥6 s in between each stimulation trial. Stimulation was performed while the FDI was at rest, as determined with EMG monitoring on an oscilloscope during the experiment.

Fig. 1.

A. placement of the grid for transcranial magnetic stimulation (TMS) mapping procedure. All points were 12 mm apart and placed based on anatomical landmarks. Mapping was done bilaterally. B: representative contra- and ipsilateral motor maps [motor evoked potential (MEP) amplitude indicated by color, blue to red in ascending order] resulting from dominant and nondominant hemisphere stimulation (strongly right-handed, Edinburgh LI = 0.78; strongly left-handed, Edinburgh LI = −0.9; mixed-handed, Edinburgh LI = −0.2). Maps are presented in a spatially compatible fashion with both the contra- and ipsilateral maps for the left hemisphere placed on the left, and likewise for the right hemisphere. The 1st map includes an overlay of the grid used in the mapping procedure. There are no relationships between the symmetry of contralateral representations and handedness although mixed handed individuals show more ipsilateral MEPs.

Behavioral data processing

Reaction times from the Poffenberger task were trimmed such that all responses <100 ms (anticipatory responses) and >3 SD from the mean reaction time (attentional lapses) were omitted. Any blocks where participants were <66% accurate in their responses were also excluded from analysis.

In the subset of participants that also underwent TMS testing, the CUD measure was subdivided to look at transfer times from one hemisphere to the other (Marzi et al. 1991). Data from four participants were excluded from this analysis due to poor performance as described in the preceding text (accuracy <66% across all blocks). DOMCUD is the CUD calculated using responses from the dominant hand (based on self-reported preference) and is indicative of transfer from the nondominant to the dominant hemisphere. The NONCUD is the CUD calculated using responses from the nondominant hand and is indicative of transfer from the dominant hemisphere to the nondominant hemisphere.

MEP data processing

EMG data were filtered on-line and digitized (10 and 500 Hz bandwidth filtering) using a Biopac MP100 system with EMG 100C amplifiers and AcqKnowledge software (BIOPAC Systems, Goleta, CA). MEP onset latency was calculated as the point at which the MEP amplitude reached 2 SD above the mean of the amplitude of the resting state muscle activity (oscillating around 0 μV) from 500 ms before and after stimulation. The peak-to-peak amplitude of the resulting MEPs was also calculated. Both contralateral and ipsilateral MEPs were defined as those having peak-to-peak amplitude of ≥15 μV, (Fig. 2) comparable to the criterion used by Wasserman and colleagues (1992). Data from one participant was dropped from analysis due to background noise that could not be filtered out.

Fig. 2.

A: representative contralateral MEP traces. The 1st 2 panels depict MEPs of ∼15 μV, whereas the 3rd panel is ∼90 μV with a similar pattern of activity. B: representative ipsilateral MEP traces in the range of 15–20 μV. These traces can clearly be differentiated from background activity even in the presence of stimulation artifact (indicated by →). C: ipsilateral traces with activity that does not reach 15 μV. While MEPs are somewhat visible in these traces, they are not clearly distinguishable from background activity. →, TMS stimulation artifact. The x axis indicates time with each trace covering 150 ms.

Statistical analyses

Relationships between handedness measures and interhemispheric communication were assessed using linear regression analyses. The number of iMEPs was investigated in relation to both handedness measures and CUD. In these cases, Poisson regression models were used. The Poisson distribution is used for evaluating count data where high-frequency events are rare (Gardner et al. 1995) as we found to be the case with the occurrence of iMEPs. Relationships with MEP amplitude and symmetry of contralateral representations were analyzed using standard linear regression analyses. To correct for multiple comparisons, we used P < 0.01 as our cut-off for statistical significance.


Behavioral measures

Table 1 shows the mean and SD of scores on the handedness and dexterity assessments by self-reported hand preference group. Scores for each hand and handedness group on the dexterity measures are presented in Supplementary Table S1.1 All of the assessments were significantly correlated with one another with the exception of the Purdue Pegboard and Grip Strength measures (Table 2). Importantly, all of our behavioral assessments were strongly correlated with the Edinburgh Handedness Inventory, indicating that preferred hand use is strongly related to upper limb laterality based on measures of dexterity. Figure 3 shows the distribution of handedness scores for all participants as well as the subset of TMS participants measured using the Edinburgh Handedness Inventory as well as the distributions of laterality scores on our dexterity measures. Figure 4 shows the distribution of scores on the laterality of dexterity measures for all participants. We over-recruited left-handed and less strongly handed participants to investigate degree of handedness and laterality effects. The Edinburgh Handedness Inventory was used for the sake of parsimony with much of the existing literature. Given the strong intercorrelations among all of the handedness and dexterity measures (Table 2), we elected to use just the two graphomotor measures of handedness (tap circles and tap squares) in subsequent analyses. In cases where results for one of these measures are presented in the absence of the other, it is because only one yielded significant results.

View this table:
Table 1.

Mean scores (SD) for self-report left- and right-handed participants on all handedness assessments

View this table:
Table 2.

Correlation coefficients of Handedness measures

Fig. 3.

A: distribution of scores on the Edinburgh Handedness Inventory for all participants. B: distribution of scores on Edinburgh Handedness Inventory for the subset of TMS participants.

Fig. 4.

A–D: distribution of scores for all participants on the laterality of dexterity measures (unimanual peg LI, grip LI, tap circles LI, tap squares LI, respectively).

Table 3 shows mean reaction times from the PP for the left and right visual field stimulus presentations blocked by responding hand. There were no significant differences in reaction time for left [F(1,46) = 0.67, P > 0.1] or right [F(1,46) = 3.3, P > 0.05] hand responses to stimuli presented in either visual field. Furthermore there were no differences for left and right-handers for responses presented in the left visual field [F(1,46) = 0.79, P > 0.05] or the right visual field [F(1,46) = 2.7, P > 0.05]. Finally, there was no significant difference in reaction time when the two visual fields were pooled across handedness groups [t(1,94) = −2.09, P = 0.04].

View this table:
Table 3.

Mean reaction times in milliseconds (SD) based on visual field of stimulus presentation and hand preference

The mean CUD across all participants was 1.74 ± 3.3 ms, which is consistent with previous findings (Marzi et al. 1991). There was no significant difference in CUD between left and right-handers [F(1,46) = 1.78, P > 0.05]. Regression analysis of the CUD and handedness laterality scores revealed a near-significant linear relationship with laterality of dexterity measured by tap circles (r = 0.33, P = 0.02). That is, those that were less strongly handed had IHTTs that were closer to zero, and left-handed individuals had the fastest IHTTs. For all other relationships, please see Supplementary Table S2.

The NONCUD and DOMCUD were strongly correlated with one another (r = −0.85, P < 0.0001) but differentially related to degree of handedness and laterality of dexterity. We found a significant positive linear relationship between NONCUD and laterality of dexterity indexed by tap circles (r = 0.64, P < 0.001; Fig. 5), indicating that left lateralized participants have faster transfer from the dominant to the nondominant hemisphere. The least lateralized participants had transfer times indicating no advantage for uncrossed responses (i.e., CUD times were close to 0). Conversely, there was a near-significant negative linear relationship between DOMCUD and tap circles (r = −0.48, P = 0.015). This relationship indicated that again those that are less lateralized have transfer times closest to zero, but in this case, those that are more strongly right lateralized have faster communication from the nondominant to the dominant hemisphere.

Fig. 5.

Correlation between the tapping circles LI score and nondominant hand crossed-uncrossed difference (NONCUD), r = 0.64, P < .001.

Contralateral motor representations

To rule out any functional abnormalities in our mixed handed participants, we evaluated whether or not there were any group differences in motor threshold or the latency of contralateral MEPs. Participants were divided into three groups including a mixed-handed group, which ranged from −40 to 40 on the Edinburgh handedness inventory. Differences in motor threshold across the two hemispheres (dominant and nondominant as determined by self-reported hand preference) and handedness group were assessed using an ANOVA. We found neither a main effect of handedness group [F(2,26) = 0.03, P > 0.5] nor hemisphere [F(1,26) = 0.92, P > 0.3]. Additionally, an ANOVA comparing handedness group and the average MEP latency for contralateral MEPs revealed no significant differences [F(2,27) = 0.44, P > 0.5]. This supports the notion that there were no functional abnormalities in our mixed-handed participants.

Maps of motor representations elicited from stimulation of the contralateral hemisphere were created for both the dominant and nondominant hand for all participants. The size of these representations was quantified by counting the number of stimulus locations resulting in at least one MEP of >15 μV peak-to-peak amplitude. Figure 1B shows both contralateral and ipsilateral maps for three representative participants—strongly right-handed, strongly left-handed, and mixed-handed. Supplementary Figure S1 includes the maps for all other individuals showing ≥10 iMEPs. We calculated the symmetry of the contralateral representations by subtracting the number of stimulation locations in the nondominant hemisphere eliciting MEPs in the contralateral hand from the number of stimulation locations in the dominant hemisphere eliciting MEPs in the contralateral hand. The contralateral representations in the dominant and nondominant hemisphere were symmetrical in the number of locations eliciting MEPs [average = 10.1 and 9.3, respectively; F(1,28) = 0.89, P > 0.05]. Further, there were also no differences in contralateral MEP amplitude across the two hemispheres [F(1,28) = 2.02, P > 0.1]. There was no significant difference in the symmetry of contralateral maps between the two handedness groups [based on self-reported handedness; t(1,27) = 1.11, P > 0.05]. There was also no significant correlation between degree of handedness and the symmetry of these representations (r = −0.21, P > 0.05). However, interestingly, when we investigated the relationship between contralateral symmetry and CUD, there was a significant negative correlation between the two (r = −0.53, P < 0.01; Fig. 6). This supports the notion that those with more symmetrical motor representations have faster CUD times, consistent with previous literature demonstrating faster CUDs with more bilateral signal processing (Saron et al. 2003).

Fig. 6.

Correlation between symmetry of contralateral motor representations (dominant hemisphere - non dominant hemisphere) and interhemispheric transfer time measured by CUD, r = −0.53, P < 0.01.

iMEP measures

To control for the possibility that iMEPs were related to motor threshold, we first evaluated whether motor threshold was correlated with number or amplitude of iMEPs. The number of iMEPs elicited from stimulation of the dominant hemisphere did not show any relationship with motor threshold (χ2 = 0.19, df = 1, P > 0.05). Likewise, when looking at iMEPs elicited by stimulating the nondominant hemisphere, there was no relationship with the corresponding motor threshold (χ2 = 1.56, df = 1, P > 0.05). Amplitude of iMEPs elicited from dominant stimulation showed a negative trend with the motor threshold of the dominant hemisphere (r = −0.51, P = 0.04). However, this relationship was in the direction opposite to what would be expected if higher motor threshold was leading to more or larger iMEPs. Finally, when examining the amplitude of iMEPs elicited from nondominant hemisphere stimulation in relation to the corresponding motor threshold, there was no significant relationship (r = 0.21, P > .05). These results indicate that ipsilateral motor activity is not due to stimulation parameters.

iMEPs were seen in 25 of 30 individuals although there was a wide range in frequency of occurrence. The number of iMEPs ranged from 0 to 85 per participant, corresponding to 0–19.7% of total stimulations (average = 2.97%). The number of iMEPs was not significantly different between the two self-reported handedness groups [t(1,27) = −1.96, P > 0.05]. The Poisson regression model relating the number of iMEPs to the absolute value of the tapping circles LI (indicative of degree of handedness) was nearly statistically significant (χ2 = 5.65, df = 1, P = 0.017), indicating that those that are less lateralized showed more iMEPs. Furthermore, when only iMEPs from stimulation of the dominant hemisphere were examined, the same pattern was observed (χ2 = 5.56, df = 1, P = 0.018). However, there was no relationship between iMEPs due to stimulation of the nondominant hemisphere and the absolute value of the tapping circles LI (χ2 = 0.33, df = 1, P > 0.05). For other relationships, see Supplementary Table S3.

Given the wide range in number of iMEPs elicited, we also performed an analysis including only individuals showing ≥10 iMEPs (10 participants). This subgroup also showed a strong relationship between the number of iMEPs and laterality of dexterity measured with the absolute value of the tapping circles LI (χ2 = 17.31, df = 1, P < 0.001; Fig. 7). Thus while iMEPs were relatively infrequent, when they did occur, they were in less strongly lateralized individuals.

Fig. 7.

Total number of ipsilateral MEPs in the subset of individuals showing ≥10 ipsilateral MEPs plotted as a function of degree of handedness indexed by the absolute value of the tapping circles LI. The Poisson regression model indicates a statistically significant relationship between these two factors, χ2 = 17.31, df = 1, P < 0.001. - - -, highlights the relationship.

To further investigate the relationship between iMEPs and handedness, we also examined iMEP amplitude. Less strongly handed individuals (as indexed by the absolute value of Edinburgh Handedness Inventory LI) did show a trend for larger iMEPs (r = −0.35, P < 0.09), although it was not statistically significant. This is consistent with the pattern previously described when looking at the number of iMEPs. Contralateral amplitude was not significantly related to degree of handedness.

We also examined the relationship between IHTT and ipsilateral motor activity. Figure 8 shows the relationships between the number of iMEPs and both DOMCUD and NONCUD. The total number of iMEPs showed a positive relationship with NONCUD such that those with slower communication from the dominant to the nondominant hemisphere had more ipsilateral motor activity. Meanwhile there was a negative relationship between the DOMCUD and the number of iMEPs—those with faster communication from the nondominant to the dominant hemisphere showed more ipsilateral motor activity. Poisson regression models indicated that both of these relationships were statistically significant (NONCUD, χ2 = 5.87, df = 1, P = 0.01; DOMCUD, χ2= 6.03, df = 1, P = 0.01). Additionally, we broke down these relationships based on hemisphere of stimulation (Fig. 9). Poisson regression models indicated a significant relationship between DOMCUD and the number of iMEPs elicited through stimulation of the dominant hemisphere (χ2 = 7.19, df = 1, P < 0.01) such that those with faster IHTT exhibited more iMEPs. The relationship with NONCUD was not statistically reliable (χ2 = 1.59, df = 1, P > 0.2) although it showed the opposite pattern as described for DOMCUD. There was a significant relationship between NONCUD and the number of iMEPs elicited through stimulation of the nondominant hemisphere (χ2 = 10.65, df = 1, P < 0.01). However, the relationship with DOMCUD was not significant (χ2 = 0.18, df = 1, P > 0.5). When MEP amplitude was broken down based on hemisphere of stimulation, there were no significant relationships (P > 0.05 in all cases).

Fig. 8.

A: total number of ipsilateral MEPs plotted as a function of NONCUD. The Poisson regression model indicates that the relationship between these 2 factors is statistically significant, χ2 = 5.87, df = 1, P = 0.01. B: total number of ipsilateral MEPs plotted as a function of dominant CUD (DOMCUD), Poisson regression, χ2 = 6.03, df = 1, P = 0.01.

Fig. 9.

Ipsilateral MEPs from dominant and nondominant hemisphere stimulation plotted as a function of DOMCUD and NONCUD. A: dominant hemisphere stimulation and DOMCUD, Poisson regression, χ2 = 7.19, df = 1, P < .01. B: nondominant hemisphere stimulation and NONCUD, Poisson regression, χ2 = 10.65, df = 1, P < 0.01.

We also computed the difference between the latency of contralateral MEPs and that of iMEPs (4.14 ± 6.57 ms). There was no significant difference between this value and CUD [1.74 ± 3.3 ms, F(1,17) = 1.95, P > 0.05], suggesting the possibility of a transcallosal mechanism underlying iMEPs. However, there was no significant correlation between these two variables (r = 0.32, P > 0.05), which is an important caveat to consider.


The present study examined the relationships between contra- and ipsilateral motor representations, degree of handedness and laterality of dexterity, and IHTT. The results indicate that less strongly lateralized individuals are more likely to exhibit iMEPs. Contralateral representations, although not related to handedness or laterality of dexterity, are related to IHTT. Furthermore, laterality of dexterity is related to interhemispheric communication based on measures of IHTT. These conclusions are supported by the following findings: 1) a greater number of iMEPs were elicited in less strongly lateralized individuals; moreover, there was a similar pattern when looking at iMEP amplitude; 2) those with faster transfer from the nondominant to the dominant hemisphere (DOMCUD) showed more ipsilateral activity, while the opposite held true for transfer from the dominant to the nondominant hemisphere (NONCUD); 3) those with the most symmetrical contralateral motor representations showed the fastest IHTT; and 4) those that are less strongly lateralized had IHTT closer to zero. These findings support the idea that degree of handedness and laterality are associated with motor cortical representations as well as interhemispheric communication. They also indicate that there may be asymmetries in communication between the hemispheres that differ with handedness and laterality of dexterity and are manifest in TMS measures.

Previous work looking at motor cortical organization has indicated a greater asymmetry in right-handed individuals. They show larger cortical volume in the dominant hemisphere (Amunts et al. 1996; Volkmann et al. 1998) and a greater volume of functional M1 activity contralateral to the dominant hand during unimanual task performance (Kim et al. 1993). Additionally, those that are less strongly handed show more functional activation ipsilateral to the moving hand (Dassonville et al. 1997). Our data indicate that these organizational differences are present at rest with less strongly lateralized individuals showing the largest amount of ipsilateral motor responses. However, our data do not support a relationship between contralateral motor cortical representations and degree of handedness, contrary to the findings of Dassonville and colleagues (1997). Given that our study employed TMS, while Dassonville and colleagues (1997) measured functional MRI activity, the difference in our findings may not be surprising. The physiological mechanisms underlying the two techniques are quite different. That is, regions that show activity when stimulated may not necessarily be recruited when performing a unimanual task. Furthermore, our findings regarding contralateral representations are in line with previous TMS studies (Cicinelli et al. 1997; Wassermann et al. 1992; Wilson et al. 1993).

Previous work reports a lack of iMEPs (Bawa et al. 2004; Carr et al. 1994; Netz et al. 1997) except for conditions in which the target muscle was contracted (Wassermann et al. 1994; Ziemann et al. 1999), although muscle contraction alters interhemispheric interactions (Nelson et al. 2009). It is therefore important to consider why we observed iMEPs in individuals at rest using only 110% of the motor threshold. Previous studies reporting a lack of iMEPs have only investigated those that are right handed (Ziemann et al. 1999). Our study included a wide range of handedness, and our data indicate that only those that are less strongly lateralized show substantial iMEPs. Given this, our results are not inconsistent with previous findings, and they underscore the importance of considering the entire handedness and laterality spectrum. It is also not clear whether these handedness effects arise during development or are a function of extended experience, particularly given evidence that practice results in increased corticospinal excitability in both hemispheres (both contra- and ipsilateral to practice) (Koeneke et al. 2006).

The level of the CNS at which ipsilateral activity is generated has been to this point unclear. It has been proposed that this occurs transcallosally due to motor overflow or else due to subcortical branching of the corticospinal tract (as suggested in Dassonville et al. 1997). In the average person, 10–15% of corticospinal projections are uncrossed, innervating ipsilateral muscles (cf. Brinkman et al. 1970; Carson 2005). It is possible that when TMS is administered, in addition to activation of crossed corticospinal projections, these uncrossed projections are activated as well resulting in ipsilateral activity. However, these ipsilateral projections predominantly innervate more proximal muscles (Brinkman and Kuypers 1973). Conversely, ipsilateral activity may also be due to interhemispheric interactions at the cortical level, causing the motor signal to cross to the other hemisphere via the corpus callosum resulting in ipsilateral activity. Although Wahl and colleagues (2007) recently demonstrated that callosal fibers connecting the hand regions of the motor cortices are related to interhemispheric inhibition, it is of note that this research only included right-handed individuals. Thus it is possible that there are differences in those that are more mixed handed resulting in different interhemispheric interactions. Furthermore, Müller, Kass-Iliyya, and Reitz (1997) demonstrated that young children between the ages of 3 and 10 regularly show ipsilateral motor activity with TMS, although this is not typically the case in those above age 10. They argue that this may be due to myelination of the corpus callosum across development, resulting in more transcallosal inhibition, although the authors do acknowledge the potential contribution of ipsilateral projections as well. Our study supports that ipsilateral activity in less strongly handed individuals is mediated by callosal mechanisms. We found that iMEPs were related to CUD and performance of a bimanual task. Furthermore, the latency difference between contra- and ipsilateral MEPs was comparable to CUD. However, the prevailing literature on this topic suggests that iMEPs are due to ipsilateral corticospinal projections (Jung and Ziemann 2006; Ziemann et al. 1999). Our data were collected with participants at rest though, and it is possible that different pathways may be involved for active versus resting states.

Alternatively, though we have demonstrated a relationship between ipsilateral motor activity and degree of laterality in conjunction with a relationship between IHTT and degree of laterality, it may be that there are differences in less lateralized individuals in both the corpus callosum and ipsilateral corticospinal projections. Our correlations between IHTT measures and ipsilateral activity may then just be an artifact of these other neuroanatomical effects in less lateralized individuals. Future studies should include additional measures of interhemispheric interactions such as paired-pulse TMS to disentangle these effects.

We parsed our behavioral measure of interhemispheric communication (CUD) in terms of transfer of information from the dominant hemisphere to the nondominant (NONCUD) and from the nondominant to the dominant (DOMCUD). However, past studies have considered this only in terms of the hand of response or visual field (Bisiacchi et al. 1994; Marzi et al. 1991) not in terms of the direction of transfer from dominant to nondominant hemisphere. Handedness has not been considered either, although Bisiacchi and colleagues (1994) suggested that it may be an important factor. Our data indicate that the time of transfer from the dominant hemisphere to the nondominant is different across the spectrum of laterality of dexterity than the time of transfer from the nondominant to dominant hemisphere. In both instances, less lateralized individuals show IHTTs close to zero. Taken together, this indicates that patterns of interhemishperic communication may differ across the handedness and laterality spectrum.

Furthermore, the differential pattern of the DOMCUD and NONCUD correlations across the handedness and laterality spectrum may reflect differing patterns of interhemispheric interactions. Left-lateralized individuals show the fastest NONCUD times. Because NONCUD is indicative of transfer from the dominant to the nondominant hemisphere, this advantage may be underlying transfer of motor planning information between the two hemispheres. The left hemisphere is thought to be the dominant hemisphere for motor planning (Haaland and Harrington 1996), thus communication from the right to the left hemisphere, and perhaps back again to the right hemisphere, may underlie the ultimate initiation of movement. DOMCUD shows the opposite pattern, such that right-lateralized individuals have faster times and this may reflect differing patterns of interhemispheric interactions. Furthermore, the finding of an asymmetry in transfer time is consistent with previous work (Marzi et al. 1991; Saron and Davidson 1989) even though we subdivided CUD based on hand dominance.

In some participants, particularly left-lateralized individuals, we found negative CUD times. Although a negative CUD may seem counterintuitive, others have previously reported negative CUDs as well (Marzi et al. 1991). EEG work by Saron and colleagues (2003) may shed some light on this finding. They noted that faster responses were associated with more bilateral activity after stimulus presentation. This is supported by our finding that those with more symmetrical contralateral motor representations had the fastest CUD. Although we see no relationship with handedness, more bilateral representations are indicative of faster IHTT. Left-lateralized individuals may engage both hemispheres for a motor task as the left hemisphere has been implicated in movement planning and preparation (Haaland and Harrington 1996), while their right hemisphere is relied on for execution of the movement. In some left-lateralized individuals, this simultaneous processing in two hemispheres while performing the PP may actually be faster than single hemisphere processing, resulting in a negative CUD.

Human handedness is the most obvious instance of motor laterality. Additionally, differences in brain structure and function are manifest across handedness groups. Sainburg (2005) views hand preference as an important underlying factor for the organization of motor control. The dynamic dominance hypothesis (Sainburg 2005) argues that the dominant limb is important for the control of movement trajectory, while the nondominant limb is important for the control of limb position. This pattern has been shown to be the opposite in left-handed individuals in studies looking at both visuomotor adaptation (Wang and Sainburg 2006) and proprioceptive matching (Goble et al. 2009). Additionally, investigations with stroke patients have also supported this theory (Schaefer et al. 2007, 2009). However, how this pattern generalizes to mixed handedness is unclear, although our data provide some insight. If in the cases of mixed handedness neither hemisphere is specialized for a particular task, then efficient communication between the hemispheres would be important for both trajectory and position related tasks. Furthermore, one might expect more ipsilateral motor activity due to this communication and perhaps involvement of the other hand to aid in efficient task performance. The faster IHTTs and greater number of iMEPs we observed in less strongly handed individuals are consistent with this prediction. The current data are also consistent with the findings of Chase and Seidler (2008) indicating that less strongly left-handed individuals show better transfer of a sensorimotor adaptation task between hands.

It is, however, important to note that when we corrected for multiple comparisons we used a statistical threshold of P < 0.01 instead of applying a strict Bonferroni correction that would have necessitated use of a P value threshold of P < 0.001. Thus future work on this topic should take a more targeted approach to investigating a limited number of variables.

Finally, a better understanding of iMEPs may be of clinical importance. Sun and colleagues (2009) recently investigated whether or not preoperative iMEPs were helpful in determining the outcome of hemispherectomy to alleviate epileptic symptoms. iMEPs were only reported in patients implying that they may serve a diagnostic function. The authors also conclude that looking at iMEPs may indeed be useful for planning procedures that avoid postoperative deficits. It is thus important to understand the conditions under which iMEPs may be elicited in healthy individuals. Our data indicate that they are more common in mixed-handed individuals, thus this may be an important factor to consider clinically.


We found that less strongly lateralized individuals show more iMEPs, and this activity is related to IHTT. However, contralateral motor cortical representations are only related to interhemispheric communication as evidenced by a correlation between the symmetry of contralateral representations and interhemispheric communication. Finally, IHTT is associated with degree of laterality of dexterity as supported by a correlation with the tapping circles LI. Our data indicate that patterns of communication between the hemispheres vary across the handedness spectrum.


No conflicts of interest, financial or otherwise, are declared by the author(s).


The authors thank A. Bosma, S. Chung, S. Gluhm, N. Goodman, T. Lodgson, and R. Trivedi for help with data collection and processing. We also gratefully acknowledge D. Goble and D. Weissman for helpful comments on previous versions of this manuscript. J. A. Bernard and R. D. Seidler are members of LIFE: The LIFE Course Evolutionary and Ontogenetic Dynamics.


  • 1 The online version of this article contains supplemental data.


View Abstract