‘Natural’ coordination between back muscles during constrained isometric trunk extension (group data)

In the ‘Natural’ condition, average EMG increased with force (Main effect of Bin: p < 0.001; Fig. 2A) and this increase did not differ between Muscles (Interactions: all p > 0.77). Average EMG was greater in Bins2–6 than Bin1, greater in Bins4–6 than Bin2, greater in Bins5–6 than Bin3, and greater in Bin6 than Bin4 (Table 1). Average EMG was significantly greater for SM than DM (Main effect: Muscle—p < 0.001, post hoc—p = 0.025) and LG (p < 0.001), with no difference between DM and LG (p = 0.17; Fig. 2A). SM EMG made a significantly greater contribution to %Total EMG than DM (Main effect: Muscle—p < 0.001, post hoc—p < 0.001) and LG (p < 0.001), with no difference between DM and LG (p = 0.14) (Fig. 3A). The %Total EMG for each muscle was consistent across Bins (Interaction: Muscle x Bin, p = 0.45), suggesting a proportional increase in EMG of each muscle with force, for the group.

Fig. 2

Group average data for average EMG and shear modulus data across Force Bins. Data are shown for average EMG from superficial (SM) and deep (DM) fibres of multifidus and longissimus (LG) during A ‘Natural’, B ‘After SM’ and C ‘After DM’, and for average shear modulus from SM and DM across Force Bins during D the ‘Natural’ condition. Mean and SD are shown. *p < 0.05 for comparison between Bins in the ‘Natural’ condition

Table 1 Statistical analysis outcomesFig. 3

Group average data for %Total EMG and %Total shear modulus across Force Bins. Data are shown for average %Total EMG from superficial (SM), and deep (DM) fibres of multifidus and longissimus (LG) during A ‘Natural’, B ‘After SM’ and C ‘After DM’ and D for average %Total shear modulus from SM and DM across Bins during the ‘Natural’ condition. Mean and SD are shown. *p < 0.05 for comparison between muscles in the ‘Natural’ condition for % =Total EMG and % =Total shear modulus and ‘After SM’ condition for %Total EMG

Change in EMG between adjacent Bins was greater for SM than LG across all ΔBins (Main effect: Muscle—p < 0.021, post hoc—p < 0.019; Fig. 4A) but did not differ between SM and DM, or DM and LG (both: p > 0.14). Change EMG was greater in ΔBin1 than in other Bins (Interaction: Condition x Bin p < 0.001, post hoc p < 0.05), with no difference between other ΔBins (all: p > 0.32).

Fig. 4

Group average data for change EMG and change shear modulus across Δ =Bins. Data are shown for average change EMG from superficial (SM), and deep (DM) fibres of multifidus and longissimus (LG) during A ‘Natural’, B ‘After SM’ and C ‘After DM’ and D for average change shear modulus from SM and DM during ‘Natural’ condition. Mean and SD are shown. *p < 0.05 for comparison between Δ Bins in the ‘Natural’ condition, #p < 0.05 for comparison between muscles in the ‘Natural’ condition

Variation in the ‘Natural’ coordination of back muscles within and between participants

Although the average mean difference in %Total EMG between repetitions for each muscle was small (SM: 14%, DM: 10%, LG: 11%; Fig. 5A), this was inconsistent across individuals. For three participants, the difference between repetitions was three times greater than for all the others. This suggests that, although many individuals used a similar coordination strategy between repetitions, for some, the strategy differed substantially.

Fig. 5

Individual participant data of the absolute between-repetition difference of the %Total electromyography (EMG) from superficial (SM), and deep (DM) fibres of multifidus, and longissimus (LG) during A ‘Natural’, B ‘After SM’ and C ‘After DM’. Pie charts depict the percentage of participants with the muscle with the highest EMG at each Force Bin in each Condition. The table displays the muscle with the highest EMG activity at each Force Bin for each participant for each task repetition at each condition. P participant, R repetition

In the ‘Natural’ condition, although group data showed greatest activation of SM and little change in any muscle as %Total EMG with increasing force, data for shown in Fig. 6 demonstrate that, for most individuals, the %Total for each muscle differed between Force Bins with patterns that varied between participants. Across Bins, SM was the muscle with the highest EMG activation for 41–59% participants, (Fig. 5), DM for 18–47% of participants and LG for 6–24%.

Fig. 6

Individual participant data for %Total electromyography (EMG) in each Force Bin. Rows depict data for each condition (‘Natural’, ‘After SM’, ‘After DM’), columns depict data for different muscles—superficial (SM), and deep (DM) fibres of multifidus, and longissimus (LG) and displays the changes in %Total that occurs between muscles in the Conditions

Muscle coordination strategy following EMG feedback of a specific muscle

When the constrained trunk extension task was repeated after practice of the task with visual feedback of SM EMG but no explicit instruction to change strategy, average DM EMG was less than in the ‘Natural’ condition (Interaction: Muscle x Condition, p < 0.003, post hoc: p < 0.02), but average SM and LG EMG were unchanged (all p > 0.39) (Fig. 2A, B). %Total EMG of each muscle was not affected by the feedback (Interaction: Muscle x Condition p < 0.001, post hoc: all p > 0.33) and, as in the ‘Natural’ condition, %Total SM EMG was greater than other muscles (post hoc: both p < 0.02), without a difference between DM and LG (post hoc: p = 0.99 Fig. 3B). Compared to the ‘Natural’ condition, the change in EMG between adjacent Bins was lower ‘After SM’ for ΔBin1 (Interaction: Condition x Bin p < 0.001, post hoc: p < 0.05).

After practice of the task with visual feedback of DM EMG, average EMG of all muscles was less than ‘Natural’ (post hoc: all p < 0.02; SM EMG also less than ‘After SM’, p < 0.001). Compared with both ‘Natural’ and ‘After SM’, %Total SM EMG decreased ‘After DM’ (Interaction: Muscle x Condition, p < 0.003, post hoc: p < 0.001), whereas %Total DM EMG increased (post hoc: p < 0.01) and %Total LG EMG did not differ between Conditions. %Total DM EMG was greater than LG ‘After DM’ (post hoc: p < 0.006).

Within- and between-participant variation in coordination of back muscles between conditionsWithin participant

Inspection of data suggests less difference in average EMG between repetitions for trials after EMG feedback than ‘Natural’ (Fig. 5A–C; ‘After SM’—9.9%, 11.7%, and 9.2%, ‘After DM’—6.2%, 7.9%, and 6.5% for SM, DM, and LG, respectively). For some participants, the same muscle had the highest EMG for four or more Bins between repetitions (‘After SM’ n = 5; ‘After DM’ n = 7), but most participants (n = 8) changed the muscle with the highest EMG between conditions.

Between participant

Inspection of individual data reveals that muscle coordination varied between participants (Fig. 5). During ‘After SM’, the muscle with the highest EMG was LG for 35–53% of participants, SM for 29–41% and DM for 18–35%. Changes in %Total EMG changes between Bins differed between participants (Fig. 6).

Following DM feedback, DM was the muscle with the greatest %Total EMG, ranging between 38 and 50% (Fig. 3). In contrast to the average group data, visual inspection indicated that %Total EMG increased across Force Bins in four participants for DM, two for SM and five for LG (Fig. 6).

Comparison between SWE and EMGComparison of group data—‘Natural’ condition

Similar to average EMG, the average shear modulus (as a proportion of shear modulus at 50% MVC) increased across Bins (Main effect, Bin: p < 0.001; Table 1; Fig. 2D), and average shear modulus was greater in Bins2–6 than Bin1, Bins4–6 were greater than Bin2, and Bin6 was greater than Bin3 (post hoc: all p < 0.01). In contrast with EMG findings, average normalised DM shear modulus was greater than SM (Main effect: Muscle; p < 0.018; Fig. 2D; note that this infers DM shear modulus was a greater proportion of the value at 50% MVC than SM, and not that the absolute shear modulus was greater). DM %Total shear modulus was greater than SM (Main effect: Muscle; p < 0.001, Fig. 3D), and did not differ between Bins (Main effect: Bin p > 0.99, Interaction: Muscle x Bin p = 0.12).

Similar to change EMG, there was a main effect of ΔBins for change shear modulus (p = 0.03). However, unlike change EMG, no difference between ΔBins for change shear modulus was detected by post hoc analysis (all p > 0.16, Fig. 4). In contrast to EMG, there was no significant difference between muscles in change shear modulus (Main effect: Muscle; p > 0.43). Consideration of individual participant data was undertaken to interpret differing observations of EMG and shear modulus.

Comparison of individual participant data in the ‘Natural’ condition

Both EMG and shear modulus showed individual differences in coordination between SM and DM. Although average SM EMG was a greater proportion of its value at MVC than DM in > 50% of participants, average normalised shear modulus SM was a greater proportion of its value at 50% MVC than DM in 33% of participants (Fig. 7). If the relationship between EMG and shear modulus was linear and consistent between muscles, it might be expected that change in EMG or shear modulus would provide similar interpretation for coordination between muscles, but this was not the case. When linear regression was fit to the relationship between EMG and shear modulus data, the relationships were generally positive and linear between 0–30% MVC (66% of participants), but the slopes and intercepts differed substantially between muscles (Table 2).

Fig. 7

Individual participant data for superficial multifidus (SM) as %Total electromyography (EMG) (left panel) and shear modulus (right panel) during the ‘Natural’ condition across Force Bins (Bin). Shapes depict different participants and are consistent between the two panels

Table 2 Correlation between shear modulus and EMG

Data for representative individual participants (Fig. 8) demonstrate different relationships between EMG and shear modulus. Although a positive linear correlation was found for most participants, the slopes and intercepts differed substantially for participants whose EMG and shear modulus show differences in interpretation of coordination of SM and DM muscles. Changes in shear modulus do not always reflect changes in EMG.

Fig. 8

Individual participant data depicting different relationships between electromyography (EMG) and shear modulus. Data are shown for three participants with distinct patterns of muscle activation across rows. A–C Columns show %Total EMG and %Total shear modulus across Bins for superficial fibres of multifidus (SM); relationship between %Total EMG and %Total shear modulus for SM; average EMG across Bins for SM and deep (DM) fibres of multifidus; average shear modulus across Bins; and the relationship between average EMG and average shear modulus for DM and SM. A This participant displays the pattern expressed in the group average data. There is a constant %Total EMG and shear modulus for SM across Bins and a corresponding strong positive correlation between them. Consistent with the uniform SM %Total EMG and shear modulus, both DM and SM average EMG and shear modulus increase with force in a similar manner, and a strong positive correlation between EMG and shear modulus. B This participant shows a relationship whereby SM %Total EMG and shear modulus show opposite trends across Bins, and corresponding negative relationship between them. Reflecting the reduction in SM %Total over Bins, the average EMG across Bins for DM increases proportionally more than SM. Average shear modulus shows the opposite trend. The relationship between average EMG and shear modulus for both SM and DM is strong and positive. C For SM, this participant shows a small decrease in % Total EMG and a small but inconsistent increase for %Total shear modulus across Bins. Consequently, there is no relationship between these variables. Average EMG across Bins displays DM increasing at a higher rate across Bins than SM and no change in average shear modulus across most Bins with a corresponding absence of correlation between these variables

Comparison of the relationship between EMG and shear modulus between repetitions and conditions

The correlation between shear modulus and EMG was consistently positive, but its slope and intercept varied between participants, muscles, repetitions and conditions (Table 2). The mean (absolute) difference in the slope of the regression line for the group did not differ between conditions (paired t tests all p > 0.15), but the [mean (SD)] percentage difference in slope varied substantially between participants for repetitions and conditions; SM–100% (122%) and DM– 316% (653%). Within-participant variability between repetitions of the task was also high (range 7–200%; Table 2). Mean within-participant variability (i.e. average difference in slope) for the conditions was [mean (SD)]: ‘Natural’—SM: 13.7 (11.6), DM: 57.7 (102.8) and ‘After SM’—SM: 26.5 (42.8), DM: 26.3 (58.6).