Exploring cephalopod object manipulation: modeling studies of neural control of Octopus suckers and squid tentacles - Frank Grasso

Frank W. Grasso1,2,3
1BioMimetic and Cognitive Robotics Lab, Brooklyn College, The City University of
New York, Brooklyn, NY
2Ecology, Evolution and Behavior Program, The Graduate Center, CUNY
3Cognition, Brain and Behavior Program, Psychology, The Graduate Center, CUNY

Cephalopods are able to perform complex grasping and manipulation of objects in
structurally complex, 3-D environments (e.g., opening jars, piling stones, grooming egg
fingers, etc.). These behaviors are comparable in complexity to those of higher
vertebrates in similar spatial configurations. The biomechanics of these behaviors in
cephalopods are realized largely through the action of muscular hydrostat (MH) systems
in which portions of muscular appendages are dynamically allocated to provide either
contractile force or stiffness where needed to suite the manipulation. This implies
substantial computational complexity in the neural systems controlling these behaviors
and makes the principles of neural control found in cephalopod MH systems topics of
great theoretical and practical interest. Serious technical challenges, however, limit our
direct study of the integrated control of these systems in behaving cephalopods. As a
parallel tool of inquiry, we developed the Artificial and Biological Soft Actuator
Manipulator Simulator (ABSAMS), a physically and physiologically constrained
computer simulation environment employed to study 3d models of MH systems (suckers, tongues, tentacles etc.) and their neural control. Muscle fibers are interlaced within the MH at many orientations but single motor neurons produce contraction only in small number of fibers. Active groups of motor units (MU) (muscle fiber groups innervated by a single motor neuron) control shape changes of the MH through cooperative and antagonistic action with other MUs. ABSAMS simulations allow us to explore the consequences of activation of single motor neurons acting alone or in coordinated groups on the shape changes of a cephalopod appendage. Evaluation of the similarity of the simulated shape changes with those of their biological counterparts provides a quantitative basis for evaluation of a model’s performance. I will report on the results from simulations of the squid tentacle strike and octopus sucker attachment as modeled in ABSAMS. One ABSAMS model quantitatively reproduced the 3d kinematics of tentacle strike observed in squid. Another model reproduced the attachment cycle of in a single octopus sucker and its relocation of a simulated object. The sucker’s sensory input was initial contact with the object’s surface, the tentacle received none. In these studies a binary (as opposed to graded) activation of individual MU’s was used. The success in reproducing the ballistic tentacle strike is unsurprising and consistent with our understanding of tentacle biomechanics. Success of binary control in the sucker model is intuitively surprising because tactile and proprioceptive feedbacks are integral components in crustacean and vertebrate object manipulation. These results support theidea that sucker object manipulation can result from the execution of a single precisely timed, multi-component motor program lacking sensory feedback. MH systems are common amongst mollusks but in cephalopods their range and functional versatility are greatly advanced compared to other mollusks such as gastropods. These results raise the possibility that a computationally simple method of achieving flexible object manipulation evolved in cephalopods as its brain and soft body co-evolved.