Spinal cord control of turtle scratching, swimming, and flexion reflex:

The turtle spinal cord is an excellent model system with which to investigate how the central nervous system selects and generates an appropriate movement for each circumstance an animal faces. The turtle uses three forms of scratching to bring its hindlimb into contact with three regions of its body surface.  (This is analogous to the two forms of scratching-elbow above hand and elbow below hand-that we use to scratch our backs, depending on the location of the itch; in both turtles and human beings, biomechanical constraints necessitate different movement strategies for scratching different regions.) The three forms of turtle scratching differ in which part of the hindlimb rubs against the body and also in when the knee extends (producing the rub) within each cycle of hip movement (Mortin et al., J. Neurophysiol. 53:1501-1516, 1985).  Analogous differences in the timing of knee extension within the hip cycle can be seen in hindlimb muscle contractions and in the electrical activity of motor neurons that activate these muscles (Robertson et al., J. Neurophysiol. 53:1517-1534, 1985).  The turtle spinal cord can produce these movements without any input from the brain. Even when muscle activity and therefore movement is prevented (using a chemical neuromuscular blocker), the spinal cord motor neurons produce action potentials in essentially these same patterns; examples of these patterns are shown in the picture below.

turtle preparation

We can then tickle a turtle in a specific location and monitor the electrical activity of individual spinal cord interneurons (which may themselves activate or inhibit motor neurons) while the turtle is producing scratch motor patterns.  Through such experiments, we can ask whether each of the three forms of scratching is selected or generated by a distinct, dedicated subset of spinal cord interneurons or whether instead individual spinal cord interneurons contribute to selecting or generating more than one form of scratching (Berkowitz, Annals of the NY Academy of Sciences 1198: 119-132, 2010).  So far, the latter appears to be the case.  Evidence has suggested that a single distributed network of spinal cord interneurons acts to select and to generate multiple forms of scratching motor patterns (Berkowitz and Stein, J. Neurosci. 14:5089-5104 & 5105-5119, 1994 and Berkowitz, J. Neurophysiol. 86:1017-1025 & 1026-1036, 2001; J. Neurophysiol. 94:4455-4470, 2005; Berkowitz et al. J. Neurophysiol. 95:2821-2831, 2006).

We can also activate forward swimming motor patterns by electrical stimulation of descending axonal pathways from the brain (Juranek and Currie, J. Neurophysiol. 83:146-155, 2000). In contrast to the three forms of scratching, the spinal cord interneurons activated during scratching and forward swimming motor patterns are only partly overlapping and there are some regional differences (Hao et al., J. Neurophysiol.112: 147-155, 2014). Many interneurons are activated during both types of motor patterns, but others are activated during scratching and inhibited during swimming (Berkowitz, J. Comp. Physiol. A 188:225-234, 2002; Berkowitz, J. Neurophysiol. 99: 2887-2901, 2008). Simultaneous scratch- and swim-evoking stimuli can produce swimming motor patterns with an increased rate, demonstrating integration of the inputs, as well as blends of the two (Hao et al., J. Neurophysiol. 106: 1766-1781, 2011); these effects are due to connections among interneurons, prior to motor neurons (Hao and Berkowitz, Frontiers in Neural Circuits, 11: 54, doi: 10.3389/fncir.2017.00054, 2017).

Flexion reflex involves some spinal cord interneurons that are also activated during scratching and swimming (Berkowitz, J. Comp. Physiol. A 188:225-234, 2002; Berkowitz et al. J. Neurophysiol. 95:2821-2831, 2006), as well as others that are selectively activated during flexion reflex (Berkowitz, J. Neurosci. 27:4634-4641, 2007). A flexion reflex stimulus that occurs during an ongoing swim motor pattern can interrupt and reset the swim rhythm, demonstrating a strong interaction between the two spinal cord circuits (Elson and Berkowitz, J. Neurosci. 36: 2819 –2826, 2016). Gentle foot stimuli that trigger flexion reflex display multi-second summation (i.e., "windup"), which is mediated partly by L-type calcium channels (Johnson et al., Frontiers in Neural Circuits 11: 83, doi: 10.3389/fncir.2017.00083, 2017).

Current experiments aim to tease out the differences between multifunctional interneurons that contribute to all these motor patterns and behaviorally specialized interneurons that contribute selectively to scratching or flexion reflex, in terms of their locations, morphologies, neurotransmitters, and whether or not they directly contact motor neuron, as well as to reveal effects of neuromodulators on innocuous flexion reflex windup and on the selection and generation of turtle spinal motor patterns generally.  A long-term challenge will be to understand how the complex and numerous connections among spinal cord neurons reliably produce an appropriate motor output for each type of sensory input.  The answer to this question may have implications for many other networks of neurons in other vertebrates.

Strong interactions between scratch and swim spinal cord networks


Hyperpolarization of scratch-specialized neurons during swim

T neurons are a subset of scratch-swim neurons

Scratch/swim neurons

Scratch-specialized neurons

T neurons, a new morphological & physiological class of spinal interneurons

Flexion-reflex specialized neurons

Broad tuning of spinal interneurons

Rhythmic activity of interneurons during two types of scratching

Rhythmic activity of interneurons during left and right motor rhythms


Back to the Ari Berkowitz home page