Temporal and spatial selectivity of hippocampal connectivity with sensorimotor cortex during individual finger movements

Cognitive control refers to brain processes involved in regulating behavior according to internal goals or plans. This study examines whether hippocampal connectivity with sensorimotor cortex during paced movements shows a pattern of spatial and temporal selectivity required for cognitive control. Functional magnetic resonance imaging activity was recorded from thirteen right-handed subjects during a paced, non-mnemonic (repetitive tapping) motor task. Direct and inverse connectivity in sensorimotor cortex were examined from psychophysiological interactions (PPI) from hippocampal seed activity during two sets of analyses: the first identified motor interactions relative to rest, whereas the second identified interactions in motor activity between fingers. Finger representations identified in a previous study were used to evaluate patterns of temporal and spatial selectivity in hippocampal connectivity. Changes in the magnitude of connectivity were identified within the sensorimotor representations of the first (index) through third (ring) fingers across time periods when each finger moved; at each finger representation, hippocampal connectivity was greatest when the represented finger was moving, reflecting temporal selectivity for the timing of finger movements. Similarly, the seeds associated with each finger representation differed in their magnitude of connectivity for adjacent finger representations, reflecting spatial selectivity for the moving finger. The patterns of spatial and temporal selectivity of connectivity during volitional movements in this study meets the criteria for cognitive control adapted from oculomotor studies.


37
To move purposefully, one or more signals within our brain direct the motor system to carry out 38 the intended movement. This is the essence of cognitive control, the process by which goals or plans influence behavior. Although various regions of prefrontal cortex have been broadly postcentral with precentral regions of sensorimotor cortex (SMC) providing sensory feedback studying cognitive control in SMC, even though such a relationship has been used effectively to connectivity is consistent with cognitive control of motor function, and describe experiments that fMRI data processing 155 SPM12 software (http://www.fil.ion.ucl.ac.uk/spm) was used to process and analyze data, applying normalization scaled the mean intensity of each brain volume to a common value to correct for 163 whole brain differences over time.

164
The representations for F1 (index finger) through F3 (ring finger) were adapted from an earlier 165 study (77). hippocampus of the normalized brain was sampled, as delimited by the aal atlas in the WFU an interaction term specified a greater effect of seed activity during movement of a single finger interaction term specified a greater effect of seed activity during movement of one finger relative to another finger (e.g., T1>T4). Each set of analyses used a constant time period for the baseline comparison (visA and T4, respectively), so that differences in connectivity between finger 181 representations or time periods would not result from different baselines. Following adjustments 182 for regional differences in timing, a regression analysis identified the magnitude of the BOLD 183 signal in SMC correlated with the PPI interaction term.

184
Seed selection 185 In this study, "functional seeds" were selected (see Burman 2019), i.e., the hippocampal voxel generating the greatest movement-related connectivity anywhere within SMC. Functional seeds were selected both for high values of direct connectivity within SMC ("MAX") and for low 188 values, which reflect inverse connectivity ("MIN"). For both direct and inverse connectivity, a identifying a finger representation for the right hand as the left SMC region where activation not be mapped, perhaps because some subjects rotated their wrist rather than moving their little 201 finger to press the key. The high-resolution activation map for each finger representation from 202 the earlier study was superimposed on a low-resolution (4mm isovoxel) image used in the 203 current study; low-resolution coordinates for each finger representation are recorded in Table 1.

204
Activation maps were flipped along the x-axis to map finger representations in the right 205 hemisphere. Because high-resolution activation did not coincide with the boundaries of the low-206 resolution voxels, a few voxels in the current study were associated with more than one finger 207 representation. With the voxels identified for each finger representation (F1-F3), the amplitude of connectivity 213 from each functional seed (S1-S3) was identified and plotted for each time period (T1-T4).

223
The absolute value of beta estimates for connectivity varied widely between subjects and 224 analyses, reflecting subject variability in several factors tangential to the analysis of interest (e.g., subject variability in spatiotemporal offsets used to best fit SMC activity to the PPI interaction changes in connectivity from the median amplitude across all conditions; the median amplitude 228 was calculated for each voxel within a finger representation, with the total range specified 229 separately for temporal and spatial selectivity analyses. A previous study showed connectivity

239
Evidence for temporal selectivity for functional seeds is shown in Figure 1 for both the L and R 240 postcentral gyrus. In the L postcentral gyrus, the PPI interaction term compared seed activity 241 during finger movements vs. the preceding visual rest condition; significant connectivity was not 242 observed when comparing seed activity during movement of one finger relative to the fourth 243 finger. In the R postcentral gyrus, the reverse was true; significant connectivity was observed 244 only when the PPI interaction term compared seed activity during movement of one finger 245 relative to activity during movement of the fourth finger.

287
For each finger representation, the spread was greatest during the period of the represented 288 finger's movement.     3). Significant connectivity in each time period (Fig 3A) 332 was present only in the finger representation for the moving finger. During T1 and T2, both bottom analyses identified connectivity associated with a specific finger; connectivity associated with the fingers was organized topographically (middle row of images). F1=yellow, F2=red, and   The topography of hippocampal connectivity with SMC was consistent with our previous study 413 (77) and those of others (78-82), which showed overlapping activation between nearby fingers.

443
In the L hemisphere, connectivity was measured relative to a baseline visual resting condition; 444 the interaction term for this connectivity analysis provides a signal simply to "move this finger," 445 suggesting direct control of an individual finger independent of others. Connectivity in the R 446 hemisphere, on the other hand, required an interaction term that explicitly signaled, "move this 447 finger but not that one," suggesting an interactive process between the control of different spatial location of the hand on the response pad was fixed in the current study, the spatial selectivity was observed between finger representations, as the magnitude of connectivity for the 477 moving finger differed from that of the surrounding fingers from the representation of the finger 478 currently moving. Thus, spatial properties of the hippocampus provided the selectivity in motor 479 response required for cognitive control.

480
The third requirement for cognitive control is temporal selectivity. The hippocampus responds 481 differentially to sequences of events that differ in order (97-100), and to the intervals between 482 stimuli within a sequence (101)(102)(103)). In the current study, cognitive awareness of the temporal 483 intervals between metronome beats provided the basis for anticipatory behavioral responses.

484
Furthermore, temporal selectivity was observed within each finger representation, such that 485 connectivity was greatest when the represented finger was moving. properties of the hippocampus. Some have suggested the hippocampus has cognitive functions reflect the varied components of episodic and long-term declarative memory (92, 93, 107).

499
Although hippocampal connectivity with SMC meets the requirements for cognitive control, 500 could the results be better explained through its other known functions?

501
Memory of the pacing interval, plus associations between the numerical onscreen display and the 502 corresponding fingers, were required for accurate performance during motor tasks in this study; 503 these modest memory requirements might arguably require hippocampal input. Neither spatial 504 nor temporal selectivity, however, is required to access this mnemonic information. Spatial 505 selectivity is unnecessary because the mapping between numerals and fingers never changed, 506 whereas temporal selectivity is unnecessary because widespread cortical rhythms (such as theta) 507 could serve to time events. The observed pattern of spatial and temporal selectivity is required 508 for cognitive control, however, to specify the finger to be moved and when. sensorimotor processing, facilitating the transformation of sensory signals into motor output 511 (108-111). This interpretation could explain why its topography and spatial selectivity were