Muscle kinematics: behavioral covariates were derived from the length of the 39 modeled muscles Chan and Moran, and their time derivatives. However, because this would result in almost 78 highly correlated covariates, we used PCA to extract 5-dimensional orthogonal basis sets for both the lengths and their derivatives.
On average, five components explained 99 and 96 percent of the total variance of lengths and length derivatives, respectively. We used repeated 5-fold cross-validation to evaluate our models of neural activity, given that the models had different numbers of parameters, P.
On each repeat, we randomly split trials into five groups folds and trained the models on four of them. We then compared the predicted firing rates from each model to the actual firing rates in that test fold, using analyses described in the following sections. Thus, if a more expressive model with more parameters performs better than a simpler model, it would suggest that the extra parameters do provide relevant information about the neural activity not accounted for by the simpler models.
To perform statistical tests on the output of repeated 5-fold cross-validation, we used a corrected resampled t-test, outlined in Ernst and Nadeau and Bengio Here, sample mean and variance are calculated as in a normal t-test, but a correction factor needs to be applied to the standard error, depending on the nature of the cross-validation.
Equation 3a-c shows a general case of this correction for R repeats of K-fold cross-validation of some analysis result d k r. The correction applied is an extra term i. Note that we performed all statistical tests within individual sessions or for individual neurons, never across sessions or monkeys. At the beginning of this project, we set out to compare three of these six models: hand-only, egocentric, and muscle kinematics.
In this analysis, we found that the muscle model performed best. As we developed this project, however, we tried the three other models to see if they could outperform the muscle kinematics model, eventually finding that the whole-arm model, built on Cartesian kinematics of the hand and elbow outperformed it.
As this appeared to be primarily due to modeling and measurement error in the muscle model see Appendix 1 , we decided to focus on the hand-only and whole-arm model. Despite only making one pairwise comparison in the main text, we chose to use a Bonferroni correction factor of 6: three for the original three pairwise comparisons and one more for each additional model we tested, which were compared against the best model at the time, and could have changed the end result of this project.
We evaluated goodness-of-fit of these models for each neuron by using a pseudo-R 2 p R 2 metric. We binned the trajectory into 16 bins, each In figures, we plotted this mean firing rate against the center-point of the bin. We calculated PDs for each neuron in each workspace and found the predicted change in PD from the contralateral workspace to the ipsilateral workspace, given each model. We compared these changes to those observed for each neuron. The values of these PD shifts are shown in Figure 7 for all neurons tuned to movements in both workspaces, averaged over all test folds.
In total, given the 20 repeats of 5-fold cross-validation, this gave us samples of the cVAF for each model in a given session. Our procedure for calculating the separability of the whole-arm kinematics was similar, simply substituting the whole-arm kinematics for the neural activity when training and testing the LDA classifier. Over the course of this project, we analyzed several different models of area 2 activity. We categorized these models into two classes based on whether they contained information about the hand or the arm in different coordinate frames.
Of these models, we picked the hand-only and whole-arm models to represent the two model classes in the main paper, as we found that the other within-class models offered little additional insight into area 2 activity.
For completeness, however, this section expands on the comparisons between within-class models. Two of our models used the kinematics of hand movement as behavioral covariates for area 2 neural activity: the hand-only model in the main paper and the egocentric model, which represents hand kinematics in a spherical coordinate frame with origin at the shoulder.
While the egocentric model, or a model like it, has been proposed as a possible coordinate frame for representation of the limb Bosco et al. Appendix 1—figure 1A and B show comparisons between the hand-only model and the egocentric model in terms of pR 2 and tuning curve correlation, as in the main paper. These comparisons show that the hand-only model tended to out-perform the egocentric model. Further, the egocentric model predicted large shifts in PD between the two workspaces Appendix 1—figure 1C that did not match up at all to the actual PD shifts.
A pR 2 comparison, as in Figure 4. B Tuning curve correlation comparison, as in Figure 6. In addition to the whole-arm model detailed in the main paper, we tested two models of area 2 activity based on biomechanics: one based on joint kinematics and the other based on musculotendon lengths. Previously, Chan and Moran used this model to analyze the joint and muscle kinematics as a monkey performs a center out task Chan and Moran, Here, we use the musculoskeletal model to predict neural activity.
Appendix 1—figure 2A and B show comparisons of pR 2 and tuning curve correlation between the whole-arm model detailed in the paper and these two biomechanical models. We found that the three models provided similar predictions, but surprisingly, the whole-arm model generally outperformed the biomechanical models. We found that neither biomechanical model could predict PD shifts as well as the whole-arm model, though the muscle model in particular appeared to perform well.
As a control for errors introduced into the muscle model by processing marker data with OpenSim, we performed the cVAF analysis on a whole-arm model where hand and elbow kinematics were derived from joint angles of the musculoskeletal model, rather than directly from the marker locations captured by the motion tracking system. We re-ran the model prediction analysis for only the muscle model, marker-derived whole-arm model, and OpenSim-based whole-arm model. Unsurprisingly, we found average cVAFs similar to those from the main analysis for the marker-derived whole-arm model 0.
This suggests that the difference in predictive capability between the muscle and whole-arm models stems at least in part from errors introduced in OpenSim modeling, rather than from the whole-arm model necessarily being the better model for area 2 neural activity. As proprioceptive signals originate in the muscles, arising from muscle spindles and Golgi tendon organs, we expected to find that the muscle model would outperform the other models.
However, there are several potential reasons why this was not so. The most important ones can be divided into two categories loosely tied to 1 errors in estimating the musclulotendon lengths, through motion tracking and musculoskeletal modeling, and 2 the fidelity of the muscle model to the actual signals sent by the proprioceptors. In the first category, the main issue is that of error propagation. The extra stages of analysis required to compute musculotendon lengths registering markers to a musculoskeletal model, performing inverse kinematics to find joint angles, and using modeled moment arms to estimate musculotendon lengths introduce errors not present when simply using the positions of markers on the arm.
As a control, we ran the whole-arm model through two of these extra steps by computing the hand and elbow positions from the joint angles of the scaled model, estimated from inverse kinematics.
The results of this analysis showed that the performance of the whole-arm model with added noise dropped to that of the muscle model, indicating that there are, in fact, errors introduced in even this portion of the processing chain.
The other potential source of error in this processing chain stems from the modeled moment arms, which might not accurately reflect those of the actual muscles. In developing their musculoskeletal model, Chan and Moran collected muscle origin and insertion point measurements from both cadaveric studies and existing literature Chan and Moran, However, due to the complexity of some joints, along with ambiguity of how the muscle wraps around bones and other surfaces, determining moment arms purely by bone and muscle geometry is a difficult problem An et al.
Because moment arms are irrelevant for determining hand and elbow kinematics, we could not subject the whole-arm model to the error introduced by this step. In addition to the questions of error propagation and musculoskeletal model accuracy is the question of whether our muscle model was truly representative of the signals sensed by the proprioceptors.
The central complication is that spindles sense the state of the intrafusal fibers in which they reside, and have a complex, nonlinear relation to the musculotendon length that we used in our muscle model. Factors like load-dependent fiber pennation angle Azizi et al. Additionally, intrafusal fibers receive motor drive from gamma motor neurons, which continuously alters muscle spindle sensitivity Loeb et al.
Altogether, this means that while the musculotendon lengths we computed provide a reasonably good approximation of what the arm is doing, they may not be a good representation of the spindle responses themselves. Spindle activity might be more accurately modeled when given enough information about the musculotendon physiology.
However, to model the effects of gamma drive, we would either have to record directly from gamma motor neurons or make assumptions of how gamma drive changes over the course of reaching. In developing models of neural activity, one must carefully consider the tradeoff between increased model complexity and the extra error introduced by propagating through the additional requisite measurement and analysis steps.
Given our data obtained by measuring the kinematics of the arm with motion tracking, it seems that the coordinate frame with which to best explain area 2 neural activity is simply the one with the most information about the arm kinematics and the fewest steps in processing.
However, this does not rule out the idea that area 2 more nearly represents a different whole-arm model that may be less abstracted from physiology, like musculotendon length or muscle spindle activity.
Still, this model comparison shows that even after proprioceptive signals reach area 2, neural activity can still be predicted well by a convergence of muscle-like signals, even though the signals have been processed by several sensory areas along the way.
One potential explanation for this is that at each stage of processing, neurons simply spatially integrate information from many neurons of the previous stage, progressively creating more complex response properties. This idea of hierarchical processing was first used to explain how features like edge detection and orientation tuning might develop within the visual system from spatial integration of the simpler photoreceptor responses Felleman and Van Essen, ; Hubel and Wiesel, ; Hubel and Wiesel, This inspired the design of deep convolutional artificial neural networks, now the state of the art in machine learning for image classification Krizhevsky et al.
Unlike previous image recognition methods, these feedforward neural networks are not designed to extract specific, human-defined features of images. Instead, intermediate layers learn to integrate spatially patterned information from earlier layers to build a library of feature detectors. In the proprioceptive system, such integration, without explicit transformation to some intermediate movement representation, might allow neurons in area 2 to serve as a general-purpose library of limb-state features, whose activity is read out in different ways for either perception or use in motor control.
Overall, our main results showed that the whole-arm model better captures firing rates and features of the neural activity than does the hand-only model. One consideration in interpreting these results is the fact that the whole-arm model is almost twice as expressive as the hand-only model, due to its greater number of parameters.
While we took care to make sure the models were not overfitting see Methods for details on cross-validation , a concern remains that any signal related to the behavior may improve the fits, simply because it provides more information. To address this concern, we would ideally compare these results with those from a model with the same number of parameters, but with behavioral signals uncorrelated with elbow kinematics, for example, kinematics of the other hand.
Unfortunately, due to experimental constraints, we only collected tracking information from the reaching arm. This model is similar to one proposed by Prud'homme and Kalaska and has the same number of parameters as the whole-arm model. While the handle forces and torques are likely correlated with the elbow kinematics, this model serves as a reasonable control to explore the particular importance of whole-arm kinematics to area 2.
Appendix 1—figure 3 shows comparisons between the whole-arm model and the hand kinematic-force model on the three metrics we used. We found that the pR 2 and the tuning curve correlation values for both models were comparable, with some neurons better described by the whole-arm model and others by the kinematic-force model.
However, we also found that the hand kinematic-force model often could not predict large changes in PD as well as the whole-arm model could Figure 7 ; Appendix 1—figure 3. In four out of eight sessions, the whole-arm model had a significantly higher cVAF than the hand kinematic-force model. In the other sessions, there was no significant difference. While the two models made similar activity predictions, the better PD shift predictions suggest that the whole-arm model is a better model for area 2 neural activity.
Same format as Appendix 1—figures 1 and 2. In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses. The authors study the explanatory power of hand- and arm- kinematic model for proprioceptive processing in Brodmann Area 2.
The authors designed two elegant paradigms to systematically vary hand and arm kinematics and examined neural responses to movement. They found that a whole arm model successfully predicted reaching-related limb state, but this model failed to characterise neuronal differences between active and passive movements.
Cortical representation of proprioception is still poorly understood, and as such, the detailed investigation offered by the authors spanning multiple tasks and converging analyses is highly timely. The reviewers were particularly appreciative of the extensive revisions carried out by the authors, providing a more nuanced understanding of the power and limitations of the whole-arm model.
Together, the reported results improve our understanding of how proprioceptive signals are represented in somatosensory cortex and open up new questions on the nature of active versus passive proprioceptive processing, with important implications for future development of proprioceptive feedback to brain machine interfaces. Thank you for submitting your article "Area 2 of primary somatosensory cortex encodes kinematics of the whole arm" for consideration by eLife.
Your article has been reviewed by three peer reviewers, including Tamar R Makin as the Reviewing Editor and Reviewer 1, and the evaluation has been overseen by Joshua Gold as the Senior Editor. The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission. In particular, the reviewers agreed that more detail relating to the somatosensory RF properties of the recorded neurons need to be considered.
In addition, it is essential that more evidence needs to be presented to clarify the unique advantages of the more complex model accounting for whole-arm position. Relatedly, the reviewers agreed that the authors need to consider alternative differences between the active and passive conditions that could have been picked up by the more complex model vicariously.
Finally, it was felt that the outlook offered by the authors regarding the "classical" model of area 2 representation was not necessarily well accepted, requiring the authors to reframe the innovation offered in the present study.
Here the authors demonstrate, over different paradigms and analyses, that activity in Brodmann area 2 can be best described when considering a combination of inputs for both hand and whole-arm position, rather than directional tuning, or muscle state. Cortical representation of proprioception is still poorly understood, and area 2 is considered a cortical proprioceptive gateway. As such, the detailed investigation offered by the authors spanning multiple tasks and converging analyses is highly timely in terms of methodology, basic understanding of proprioception and BMI applications.
Moreover, by demonstrating that a given population in SI factors information of both hand and arm parameters, it challenges the canonical 'homunculus' account of SI, by which specific body parts are represented in separation to each other. While there was a lot to like in the paper, there were also some important confounds, some of which already brought up by the authors, that the reviewers felt needed to be addressed more comprehensively.
Reviewer 1 was questioning whether the extrinsic hand model is a scarecrow model — after all, the authors are recording activity from the S1 arm area. So it is not a great wonder that this area is more tuned to arm inputs than to the hand? In other words, perhaps if the authors recorded from the hand SI area, they might find greater invariance to arm position?
They also asked for a confirmation that the authors are recording from area 2 and not area 5 which is already considered to have a more complex higher-order representation? This is especially important for the population analysis that could be driven by a subset of neurons with the most relevant information.
Figure 19 of Pons et al. Both Reviewers asked that the authors demonstrated the implant location more accurately the cartoon in Figure 1 leaves a lot to be desired. From the Materials and methods it appears that the authors did some classic assessment of RF properties, and all three reviewers wanted to learn more about the tuning properties, especially with respect to the hand and arm. Similarly, Reviewer 2 pointed that considering the broad receptive fields of area 2 e. Pons et al. For example, does the neuron representing elbow kinematics possess the passive receptive fields of the hand?
This would suggest the convergence of high-level representation onto area 2. The main finding is the contribution of elbow's kinematics for generating the neuronal firing patterns observed during movement. But as the authors discuss, the performance of the model should be improved by adding multiple and different potential covarying inputs. Reviewer 2 suggested that we could expect comparable improvement if other parameters were fitted, like grip force, shoulder or trunk position, or even the kinematics of another arm.
Also, during the second behavioural task, were kinematic reference frames better at accounting for the neuronal firing during specific phases of the movement? As such, the current model comparison leaves a lot to be desired, and the reviewers felt that the main argument offered by the authors was not sufficiently conclusive.
Reviewer 2 kindly offered potential practical solutions to improve the present evidence. For example, the authors could add another model evaluating solely the contribution of elbow and shoulder position.
This is comparable with the model of "extrinsic" and "egocentric" in light of the model's complexity. By finding higher performance with the elbow model, for example, the authors could suggest a considerable contribution of proximal joints. It looks that the performance of two models is comparable see sec in Figure 5C during some movement but not the other sec in the same figure.
How was the kinematics of hand and elbow during those periods? How did they covary? Adding a 'proper' control — shuffled noise and kinematics that covaried with endpoint but less important than elbow e. Reviewer 1 pointed that from the example in Figure 3A it seems that activity no. If so, could this account for some of the key differences, e. If so, could this be explained by trivial task differences e. More interestingly, could this 'main effect' of activity that I am postulating here be driven by motor inputs?
If so, it might have a unique time-stamp? This was also echoed in Reviewer 2's point that the major difference between active and passive movement is the existence of "corollary discharge" or "efference copy", so the complex model may actually decode this in addition to the endpoint. Reviewer 2 also pointed that this result may suggest that the extent of sensory cancellation by ego-motion is different between the control of endpoint or more proximal joints.
If so, can the authors link this implication with the proposal of whole arm representation in area 2? First, the portrayal of area 2 as an area where 'neural activity is related simply to the movement of the hand and interaction forces it encounters during movement, similar to our conscious experience of proprioception' seems over-stated.
The authors cite several previous studies who suggested movement direction encoding in S1 in monkeys most prominently by the Miller group , but the reviewers felt that even when considering these citations, this statement is far from the consensus for how proprioception is represented in area 2.
Even though, historically, endpoint kinematics was thought to be the guiding principle in motor cortex, this preconception is not widely held in area 2.
As such, the reviewers agreed that the null hypothesis of this paper Abstract is not strong enough, and requires rethinking. The manuscript should therefore be substantially re-written to highlight the novel contributions of the empirical work in characterising the neuronal properties in this area, independently of the end-point model. First, we would like to clarify that the hand kinematics we refer to are the kinematics of translation of the hand through space, not motion of the fingers.
Whereas we would expect to find the latter in the hand area of area 2, we likely would not find the former.
This hand-only model has typically been used when studying the arm representation of area 2. As with the classic reaching studies of M1 beginning with Georgopoulos, its appeal is its reasonable accuracy despite its simplicity. Our main purpose is to determine how much additional information we might learn about area 2 activity by considering the movement of the whole arm. We have restructured the paper to reflect this purpose. To achieve this, we first swapped the order in which we describe the experiments, showing first that we gain explanatory power during the two-workspace task by considering the movements of the whole arm.
We have now added to the manuscript more detail on the implant location, along with results from sensory mappings on the arrays now Figure 1. In our mappings, we found a rough gradient across the arrays of Monkeys C and H corresponding to a somatotopy from proximal to distal receptive fields on the arm, which agrees with Figure 19 of Pons et al. Further, we found a mixture of cutaneous and deep receptive fields across our arrays, which matches descriptions of area 2 from previous anatomical studies Seelke et al.
Unfortunately, we did not perform sensory mappings around the same time periods as we collected the behavioral data. As such, we cannot confidently match up neural receptive field properties with model performance. However, there was no link between model performance in the two-workspace experiment and electrode location on the array, suggesting that the whole-arm model predictions of area 2 activity did not depend on the exact receptive field of the recorded neurons.
It is not clear to us what it would mean for a neuron with elbow responses to have a hand receptive field. In any case, neurons did not typically have this kind of extended, combined receptive field.
We thank the reviewers for their many different suggestions for model comparisons that could make our arguments stronger. We have chosen two additional models to add to the supplementary information to strengthen our case that the whole-arm kinematics do matter, rather than just being epiphenominal. These model comparisons help to demonstrate the particular covariates that matter most for predicting neural activity in area 2.
These include an extension of the hand-only model with added forces and torques acting on the manipulandum handle. While the two models had similar performance in pR 2 and tuning curve correlation, the force-augmented model could not predict PD shifts as well as the whole-arm model.
We have also added a joint kinematic model, which relates neural activity to the angles and angular velocities of the 7 degrees of freedom in the arm, with even more parameters than the whole-arm model. Like the muscle model, this joint model performed worse than the whole-arm model. While the amount of information added by a movement-related signal could be reduced by shuffling it, it is not clear to us to what end that would be done.
We agree that it would be interesting to look at how hand and elbow kinematics covary at points when the models make similar or different predictions. Unfortunately, because our task involved reaching to random targets, where trials were never repeated, such a covariance analysis would be difficult to perform and subsequently interpret.
However, one might imagine a follow-up experiment in which we identify sequences of targets that result in either similar or different predictions and have the monkey repeat these sequences at random, in order to adequately analyze the covariance of hand and elbow in these different cases. The new section demonstrates that whole-arm kinematics are not always sufficient to explain area 2 activity, which also acts as a control for the analyses of the two-workspace section.
In doing so, we found that while the whole-arm model could explain separability well, it was not actually capturing a consistent relationship between neural activity and arm kinematics for many of the neurons. This finding suggests that the separation in area 2 activity between active and passive movements cannot be explained purely by arm kinematics. Describe how propioception is the sense of the position of parts of our body in a three dimensional space.
Proprioception is the sense of the relative position of neighboring parts of the body and the strength of effort being employed in movement. It is distinguished from exteroception, perception of the outside world, and interoception, perception of pain, hunger, and the movement of internal organs, etc.
The initiation of proprioception is the activation of a proprioreceptor in the periphery. The proprioceptive sense is believed to be composed of information from sensory neurons located in the inner ear motion and orientation and in the stretch receptors located in the muscles and the joint-supporting ligaments stance.
Conscious proprioception is communicated by the posterior dorsal column—medial lemniscus pathway to the cerebrum. Unconscious proprioception is communicated primarily via the dorsal and ventral spinocerebellar tracts to the cerebellum. An unconscious reaction is seen in the human proprioceptive reflex, or Law of Righting.
In the event that the body tilts in any direction, the person will cock their head back to level the eyes against the horizon. This is seen even in infants as soon as they gain control of their neck muscles. This control comes from the cerebellum, the part of the brain that affects balance. Muscle spindles are sensory receptors within the belly of a muscle that primarily detect changes in the length of a muscle. They convey length information to the central nervous system via sensory neurons.
This information can be processed by the brain to determine the position of body parts. The responses of muscle spindles to changes in length also play an important role in regulating the contraction of muscles.
Muscle spindle : Mammalian muscle spindle showing typical position in a muscle left , neuronal connections in spinal cord middle , and expanded schematic right. The spindle is a stretch receptor with its own motor supply consisting of several intrafusal muscle fibers.
The sensory endings of a primary group Ia afferent and a secondary group II afferent coil around the non-contractile central portions of the intrafusal fibers. The Golgi organ also called Golgi tendon organ, tendon organ, neurotendinous organ or neurotendinous spindle is a proprioceptive sensory receptor organ that is located at the insertion of skeletal muscle fibers onto the tendons of skeletal muscle.
It provides the sensory component of the Golgi tendon reflex. The Golgi organ should not be confused with the Golgi apparatus—an organelle in the eukaryotic cell —or the Golgi stain, which is a histologic stain for neuron cell bodies.
Golgi tendon organ : The Golgi tendon organ contributes to the Golgi tendon reflex and provides proprioceptive information about joint position. The Golgi tendon reflex is a normal component of the reflex arc of the peripheral nervous system. In a Golgi tendon reflex, skeletal muscle contraction causes the agonist muscle to simultaneously lengthen and relax. This reflex is also called the inverse myotatic reflex, because it is the inverse of the stretch reflex.
Although muscle tension is increasing during the contraction, alpha motor neurons in the spinal cord that supply the muscle are inhibited. However, antagonistic muscles are activated. The somatosensory pathway is composed of three neurons located in the dorsal root ganglion, the spinal cord, and the thalamus. A somatosensory pathway will typically have three long neurons: primary, secondary, and tertiary. The first always has its cell body in the dorsal root ganglion of the spinal nerve.
Dorsal root ganglion : Sensory nerves of a dorsal root ganglion are depicted entering the spinal cord. The axons of many of these neurons terminate in the thalamus, and others terminate in the reticular activating system or the cerebellum. In the case of touch and certain types of pain, the third neuron has its cell body in the ventral posterior nucleus of the thalamus and ends in the postcentral gyrus of the parietal lobe. In the periphery, the somatosensory system detects various stimuli by sensory receptors, such as by mechanoreceptors for tactile sensation and nociceptors for pain sensation.
The sensory information touch, pain, temperature, etc. Generally, there is a correlation between the type of sensory modality detected and the type of afferent neuron involved. For example, slow, thin, unmyelinated neurons conduct pain whereas faster, thicker, myelinated neurons conduct casual touch.
In the spinal cord, the somatosensory system includes ascending pathways from the body to the brain. One major target within the brain is the postcentral gyrus in the cerebral cortex. This is the target for neurons of the dorsal column—medial lemniscal pathway and the ventral spinothalamic pathway.
Note that many ascending somatosensory pathways include synapses in either the thalamus or the reticular formation before they reach the cortex. Other ascending pathways, particularly those involved with control of posture, are projected to the cerebellum, including the ventral and dorsal spinocerebellar tracts.
Another important target for afferent somatosensory neurons that enter the spinal cord are those neurons involved with local segmental reflexes. Spinal nerve : The formation of the spinal nerve from the dorsal and ventral roots. The primary somatosensory area in the human cortex is located in the postcentral gyrus of the parietal lobe. This is the main sensory receptive area for the sense of touch. Like other sensory areas, there is a map of sensory space called a homunculus at this location.
Areas of this part of the human brain map to certain areas of the body, dependent on the amount or importance of somatosensory input from that area. For example, there is a large area of cortex devoted to sensation in the hands, while the back has a much smaller area. Somatosensory information involved with proprioception and posture also target an entirely different part of the brain, the cerebellum. This is a pictorial representation of the anatomical divisions of the primary motor cortex and the primary somatosensory cortex; namely, the portion of the human brain directly responsible for the movement and exchange of sensory and motor information of the body.
The thalamus is a midline symmetrical structure within the brain of vertebrates including humans; it is situated between the cerebral cortex and midbrain, and surrounds the third ventricle. Its function includes relaying sensory and motor signals to the cerebral cortex, along with the regulation of consciousness, sleep, and alertness.
Thalamic nuclei : The ventral posterolateral nucleus receives sensory information from the body. The cortical sensory homunculus is located in the postcentral gyrus and provides a representation of the body to the brain. A cortical homunculus is a pictorial representation of the anatomical divisions of the primary motor cortex and the primary somatosensory cortex; it is the portion of the human brain directly responsible for the movement and exchange of sensory and motor information of the body.
There are two types of homunculus: sensory and motor. Each one shows a representation of how much of its respective cortex innervates certain body parts. The primary somesthetic cortex sensory pertains to the signals within the postcentral gyrus coming from the thalamus, and the primary motor cortex pertains to signals within the precentral gyrus coming from the premotor area of the frontal lobes. These are then transmitted from the gyri to the brain stem and spinal cord via corresponding sensory or motor nerves.
The reason for the distorted appearance of the homunculus is that the amount of cerebral tissue or cortex devoted to a given body region is proportional to how richly innervated that region is, not to its size.
The homunculus is like an upside-down sensory or motor map of the contralateral side of the body. However, pain sensation often returns, albeit in a different form, following the surgical section of the spinothalamic tract. He also exhibits loss of discriminative touch and proprioception in a corresponding area on the right side of his body. Symptoms: The patient exhibits a loss in voluntary control of the right leg. He also reports loss of sensation in both feet Figure 5.
Physical examination determines that there are losses of pain and temperature sensations involving the left half of his body starting just below the left nipple and extending down to include his left foot. There are also loss of vibration and position sensations and poor localization of tactile stimuli on the right side of his body starting just below the right nipple and extending down to include his right foot.
The Romberg test is positive i. You conclude that the somatosensory losses in his body Figure 5. Neither a vibrating tuning fork applied to the right foot nor a pin prick applied to the left foot result in the appropriate sensations. Press FOOT to view the course of action potentials generated in response to the tuning fork on the right foot and a pin prick to the left foot. Press HAND to view the course of action potentials generated in response to the tuning fork on the right hand and a pin prick to the left hand.
The symptoms are bilateral - with discriminative touch and proprioception lost on the ipsilesional side and pain and temperature affected on the contralesional side. Hemisection of the Spinal Cord. The symptoms resulting from hemisection of the spinal cord i. For example, if the right spinal cord is sectioned, say at T5, the motor effect is on the right side starting at the chest and extending down to and including the right leg and foot.
Because spinal cord hemisection interrupts both the posterior column and spinothalamic tracts, there will be sensory losses that are bilateral: ipsilesional for the posterior column discriminative touch and proprioception and contralesional for the spinothalamic tracts pain and temperature. The patient suffers from loss of pain and temperature sensations that wrap around his body at his waist.
Symptoms: The patient exhibits loss of pain and temperature sensations that are bilateral and limited to his waist area i. While pain sensation is diminished around the waist, it is normal above and below the waist.
Discriminative touch, vibration and position senses are normal in the waist area and for the rest of the body and face. Pin pricks applied anywhere around the waist do not produce well-localized, sharp pain sensations. Press WAIST to view the course of action potentials generated in response to a pin prick to the right and left side of the body at the waist.
Pin pricks applied to the feet produce well-localized sensations of sharp pain. Press FOOT to view the course of action potentials generated in response to a pin prick to the right and left feet.
Pin pricks applied to the hands produce well-localized sensations of sharp pain. Press HAND to view the course of action potentials generated in response to a pin prick to the right and left hands.
In syringomyelia , there are cysts that form within the spinal cord near the central canal Figure 5. As the cyst grows, it first compresses and then destroys the decussating fibers in the anterior white commissure. Many of these fibers belong to the spinothalamic tracts and the resulting sensory loss involves pain and temperature sensation bilaterally and segmentally. The bilateral loss is described to form a belt or girdle pattern - if the damage involves the lower thoracic segments, and does not involve sensation below and above the cyst i.
Brain Stem. Trauma, stroke, multiple sclerosis a disease of myelin , and brain tumors are the major causes of brain stem lesions. The location of the lesion site can often be deduced by the loss in cranial nerve function.
The patient suffers from a decrease in pain and temperature sensations involving the left side his body and the right side of his face. Symptoms: The patient exhibits decrease in pain and temperature sensations that involve the left side of his body and right side of his face Figure 5. Discriminative touch, vibration and position senses are normal in these areas. Touch, vibration, position, temperature, and pain sensations are normal for the rest of the body and face. Whereas neurons of the spinal trigeminal pathway STP process pain, temperature and crude touch information from the face.
Pin pricks into the right side of the face and the left hand do not produce well-localized, sharp pain sensations. The vibration of a tuning fork applied to the right jaw and left hand, as well as manipulation of the jaw and fingers of the left hand produce normal vibration and proprioceptive sensations. Press TOUCH to view the course of action potentials generated in response to a vibrating tuning fork applied to the right side of the face and the left hand.
Notice that the medial lemniscus and ventral trigeminal lemniscus, which are located in the anteromedial medulla, have been spared by this infarct. Wallenberg's Syndrome. In the medulla, both the spinothalamic tracts and the spinal trigeminal tracts are located posteriorly in the area that normally receives blood via branches of the posterior inferior cerebella artery PICA Figure 5.
Consequently, an obstruction of the PICA blood supply to the medulla will result in analgesia and thermo-anesthesia of the contralesional body spinothalamic tracts and of the ipsilesional face spinal trigeminal tract. Branches of the anterior spinal and vertebral arteries supply more anterior areas of the upper medulla. Therefore, an infarct involving the PICA blood supply will not affect the medial lemniscus or ventral trigeminal lemniscus.
Consequently, discriminative touch and proprioception from the body and pain, temperature and crude touch in the contralesional half of the face will not be affected with an infarct involving PICA. The descending spinal trigeminal tract and nucleus and the ascending spinothalamic tract would be damaged, whereas the medial lemniscus and ventral trigeminal lemniscus would be spared.
Above the level of the pons Figure 5. Somatosensory Cortex. The sensory loss from head trauma or stroke that damages the somatosensory cortex will. The patient suffers from deficits in discriminative touch and proprioceptive sensations involving the right side of his body and face. Tactile and pain sensations are also poorly localized on his right side. He has difficulty walking and controlling his right arm and hand and the right side of his face.
Symptoms: The patient exhibits deficits in fine motor control and in discriminative touch and proprioception on the right side of his body and face Figure 5. He has problems manipulating and identifying objects placed in his right hand stereognosis. He is unable to identify letters or numbers written on the skin of the right face and the palm of his right hand graphesthesia.
He also has difficulty in judging weight differences baragnosis and cannot appreciate textures with his right hand. He is unable to detect the passive movement of his right foot and the fingers of his right hand. Compared with the left side of his body pain sensations are not as sharp, well defined or easily localized on the right side of his body. Touch, vibration, position, thermal, and pain sensations are normal for the rest of the body and face.
The patient has difficulty walking and the Romberg test is positive. The neurons of the spinal trigeminal pathway STP process all pain, temperature and crude touch information from the face. The vibration of a tuning fork applied to the right jaw or right hand, as well as manipulation of the right foot, produce no vibration or proprioceptive sensations.
Press TOUCH to view the course of action potentials generated in response to a vibrating tuning fork applied to the right jaw and the right hand. Pinching the right cheek or right hand produce pain sensations. Press PINCH to view the course of action potentials generated in response to pinching the right side of the face and the right hand. Hemorrhage limited to somatosensory parietal areas produces contralesional astereognosis, baragnosis, and losses in the ability to discriminate object size and texture.
Also decreased or lost on the contralesional side of the body are the ability to discriminate position and movement of body parts and the control of fine movements.
The hemorrhage would not produce a total loss of pain sensation as other cortical areas are also involved in the perception of painful stimuli. For example, the cingulate gyrus in the frontal lobe and part of the insular cortex appear to be involved in the perception of, and emotional reaction to, painful stimuli Figure 5.
The thalamic neurons of the spinothalamic pathways and spinal trigeminal pathway that are involved in processing pain information send their axons to the cingulate gyrus and insular cortex. Consequently, damage limited to the somatosensory parietal cortex will not result in the loss of all pain sensation. From this chapter, you should have learned how the somatosensory system is organized from the skin, muscles and joints to the cortex.
Information coded and carried by thousands of spinal cord and cranial ganglion cells are distributed to millions of cortical neurons in the parietal lobe. The perceptions of coherent somatosensory stimuli and body image are recomposed out of these fragments of information by the simultaneous activation of large areas of cortex.
You have learned how to use the somatotopic organization and the modality specificity of the different somatosensory pathways to determine the location and extent of damage to the somatosensory structures. The pars opercularis of the parietal lobe forms the "upper lip" of the lateral fissure and contains both visceral sensory cortex and the secondary somatosensory cortex.
The insula is the site of the gustatory cortex and more visceral cortex. The posterior parietal lobe is located caudal to the postcentral gyrus and serves as the somatosensory association cortex.
The buttock, leg, foot, and genitals are represented in the posterior paracentral lobe, which is located on the medial aspect of the cerebral hemisphere. This is incorrect, as the posterior funiculus contains first order afferents of the medial lemniscal pathway, which processes discriminative touch and proprioception.
The neospinothalamic pathway processes sharp pain sensation from the body and the second order axons of this pathway are in the lateral and anterior funiculi the spinothalamic tract. This is incorrect, as the first order medial lemniscal afferents do not decussate.
Consequently, the sensory loss is ipsilesional when these afferents are destroyed. This is incorrect, as the medial lemniscal first order afferents innervating the arm and hand enter the spinal cord posterior funiculus via posterior roots above T6. The lesion produces a positive Rhomberg sign as there is a loss of proprioception in the ipsilesional leg and the patient is unable to maintain his balance when his eyes are closed and his feet are close together.
0コメント