In psychomotor habilitation and rehabilitation by newborn and children, and by adult after head injury, encephalitis, cerebral infarctus, coma, ect.

Forgive the scant response, have only a minute at work, can fill in more later if needed:

Never heard the terminology motor primitives or synergies (thanks for the link!), but there is a load of excellent work on modular motor control in the spine using 'central pattern generators' as the organizing term. 

Sten Grillner's lab is the go-to, this review of his is the best in my opinion: 10.1016/j.neuron.2006.11.008

They do a lot of modeling work that you might be able to build on, most have spinal models that are undergoing descending modulation, for example this paper on reticulospinal modulation of steering:

.www.pnas.org/cgi/doi/10.1073/pnas.1401459111

or modular recruitment of motor pools depending on summation of descending command from David McLean:

http://dx.doi.org/10.1016/j.neuron.2014.06.021

More specific work on spinal CPG's and motor pools can be found from McCrea:

10.1111/j.1469-7793.2001.0041b.x

and Rossignol:

10.1152/physrev.00028.2005.

Dear Dr. Saunders,

thank you for your reply. I will read the papers you have suggested to me.

Anyway, to my knowledge, CPG function as self-sustaining oscillators for cyclical behaviors such as locomotion. I think that it is a bit different from a reaching movement.

The MSE (motor synergy encoder) found by Levine and colleagues should be different s from CPG

Dear prof Ilankovic , thank you fro your reply. Could you please explain me better?

Dear Antionio, You can see our articles about Diagnostic (assesment) and Rehabilitation in  psychotic residual states and You van find the principal explanation of our thesis about psychomotor regression. For rehabilitation is necessery to know very well the psychomotor development, assesment of  the nivou of regression and than the upgrading on tthe functional nivou according the priciples and phases of normal psychomotor development.

HI, please refer to these works, 

these are more from a rehabilitation (robotics) perse, but will give some idea about computational models:

1)

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3630334/

2) http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3749429/

Dear Dr. Budhota, thank you for your  references, they are very useful!

Dear Antonio,

Apart from the works you already mentioned, I find very useful the following:

Article Matrix factorization algorithms for the identification of mu...

Article Motor primitives in vertebrates and invertebrates

Article Coordination and localization in spinal motor systems

Article The supraspinal control of mammalian locomotion

Article The neural origin of muscle synergies

I know I'm late to the party on this, but I thought to provide an answer anyways.

The answer to this question is neither straightforward nor consistent across classes of movements.  This is because a primitive represents a movement class, a synergy does not.  A synergy represents a strategy to fulfill a movement class.  More specifically, it specifies the control signal to realize a type of primitive.  I try to explain a little more below.

By and large, movement primitives can be divided into several class (a la Hogan and Sternad).  Minimally, you could have discrete primitives, impedance primitives, or  rhythmic primitives.  What makes these primitives is that, presumably, any of them could be combined into a more complex movement.   

I think the nature of your question is this:  how might the central/ and periperihal nervous system represent and institute such a primitive.  That is, in what sense is it constrained by interneuron regulation by the CNS.  On the other side of this coin, in what context are these synergies due to biomechanical or physiological constraints.  

However, talking about synergies and primitives as one in the same is difficult.  Let's take a simple reaching movement to a target.  Let's ignore the shoulder, and only consider the reach as requiring torques around a two-planar link arm. We assume two agonist-anatognist pairs of muscle, one for each link.  Let's assume the CNS (e.g., the motor cortex) represents the required transformation to endpoint force (at the hand) as a trajectory.  Let's further assume this is represented as a motor primitive in terms of a simple equation of motion as follows.  F(t) = -kX(t)+bX(t)' +u(t).  A simple forced damped mass spring.  The system simply needs to translate the desired force pattern into a control signal to follow some desired path (a difficult problem I won't discuss).  Consider the following two ways this could happen: (1). the CNS represents the necessary activity to each muscle independently; therefore, not really fulfilling the requirement of synergy. (2). the CNS represents the activity as modulating something at the level of interneurons, which individually might project to multiple motor neurons; this fulfills the synergy condition.  In both scenarios, though, the same motor primitive is used.  Thus, the distinction between synergy and primitive, I believe, is required. 

Dear Justin,

thank you a lot for your answer. You are not late to the party but i need some explaination from you.

If i'm not in error, you want to say that:

1) motor primitives and synergy are two different stuffs;

2) synergies are a way to use motor primitives for executing movements

I agree with you. I use motor primitives as synonym og motor synergy because in some paper it seems that authors use them as the sam things.

Anyway, in your opinion how can a motor primitive be implemented by a physical point of view? How a motor primitive is made? Can you suggest me any paper about it?

Thank you

Antonio

Dear Alessandro, thank you for your suggestions. I read almost all of them but one of them is new for me, Thank you for your support

These are a few that are more or less similar to that posted by Alessandro. 

The answer to this is not clear cut, but let me try.

1). I am saying they are potentially two different things, depending on how you define a synergy, primitive, or your theoretical bent.  For example, the group of Bizzi et al. would basically say the spinal system is setup to activate motor primitives. Essentially, the network topology via interneurons, motorneurons, just needs the proper activation patter (pre-motor drive) provided by the CNS. 

Caveat: but a synergy is loosely defined as building blocks (read as primitives) that are temporarily assembled to complete an action. Ergo, combining primitives might constitute a synergy.  However, a primitive in its own right might be composed as a synergy. 

Part of understanding the answer requires consideration of what properties the central nervous system represents when recalling and generating movement initiating commands.  Does the motor system represents forces in a precision grip, or maybe amplitude and direction in reaching? Then what of this information is passed down to the spinal system.

2.) the answer to this is yes.  A synergy constitutes a means to drive a primitive.

A note on the properties of synergies is notion of task-relevant and irrelevant variability.  That is, this notion is also psychological; see anything on uncontrolled manifold.  Synergies should have the property that they are task selective; e.g., reaching for a stable versus an unstable cup leads to different motor commands for endpoint stability reasons (not spilling the cup).  Motor primitives themselves are agnostic to these environmental properties. 

Also, this selection of articles is by no means exhaustive.

Article Representation of Muscle Synergies in the Primate Brain

Article Spinal Interneurons Facilitate Coactivation of Hand Muscles ...

Article A Neural Basis for Motor Primitives in the Spinal Cord

Article Dynamic Primitives of Motor Behavior