Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 4, e07436

Neural Population Dynamics in Human Motor Cortex During Movements in People With ALS


Neural Population Dynamics in Human Motor Cortex During Movements in People With ALS

Chethan Pandarinath et al. Elife.


The prevailing view of motor cortex holds that motor cortical neural activity represents muscle or movement parameters. However, recent studies in non-human primates have shown that neural activity does not simply represent muscle or movement parameters; instead, its temporal structure is well-described by a dynamical system where activity during movement evolves lawfully from an initial pre-movement state. In this study, we analyze neuronal ensemble activity in motor cortex in two clinical trial participants diagnosed with Amyotrophic Lateral Sclerosis (ALS). We find that activity in human motor cortex has similar dynamical structure to that of non-human primates, indicating that human motor cortex contains a similar underlying dynamical system for movement generation.

Trial registration: NCT00912041.

Keywords: computational neuroscience; dynamical systems; human; human biology; medicine; motor control; motor cortex; neuroscience.

Conflict of interest statement

The authors declare that no competing interests exist.


Figure 1.
Figure 1.. Neural population responses show rotational activity during movement epochs.
(A) Projections of the neural population response onto the first jPCA plane for a monkey during an arm-reaching task (monkey N, 108 conditions; adapted from Churchland et al., 2012). Each trace plots the first 200 ms of activity during the movement epoch for a given condition. Traces are colored based on the preparatory state projection onto jPC1. a.u., arbitrary units. (B) Projections for participant T6 during an 8-target center-out task controlled by index finger movements on a computer touchpad. Each trace plots the 250 ms of activity during the movement epoch (‘Materials and methods’) for a given condition. (C) Same as (B), for participant T7. Video 1 shows the evolution of the neural state over time for each participant. DOI:
Figure 2.
Figure 2.. Rotational dynamics are not a by-product of the jPCA analysis method.
For each data set, neural responses were shuffled in a manner that preserved the complexity of individual response patterns on each electrode, but disrupted the structure of the data across electrodes. For each channel, the pattern of activity during the movement epoch was inverted for half the conditions (chosen at random). The inversion was performed around the initial time point, so that continuity with pre-movement activity was preserved. Performing jPCA on the shuffled responses did not reveal consistent rotational structure. (A, top) Projection of the population responses onto the first jPCA plane for a single shuffled trial (participant T6). (bottom) Fraction of variance of the change in neural state (6-D) explained by rotational activity for the original data set (brown) vs the shuffled data sets (blue). Error bar represents the standard deviation across 300 shuffle trials. (B) Same as (A), for participant T7. Two additional shuffle control analyses are presented in Figure 2—figure supplement 1 and Figure 2—figure supplement 2. DOI:
Figure 2—figure supplement 1.
Figure 2—figure supplement 1.. Results of the second shuffle control analysis.
This shuffle control was similar to the first, but the patterns of activity during the movement epoch were inverted for all conditions. (A) Projections of the neural population state for each data set after shuffling are shown. (B) The fraction of variance captured by the rotational model before and after shuffling. This manipulation would not be expected to remove all rotational structure, as the structure should merely be sign-inverted. However, this manipulation should remove any consistent relationship between the pre-movement activity and the phase of subsequent oscillations. Thus, this control is expected to disrupt the relationship between the rotational phase and its initial state. DOI:
Figure 2—figure supplement 2.
Figure 2—figure supplement 2.. Results of the third shuffle control analysis.
In this shuffle control, activity during the movement epoch from one condition was randomly reassigned onto the pre-movement activity from another. The firing rate at the beginning of the movement epoch was appended to the firing rate at the end of the pre-movement period, to preserve continuity. For a given shuffle trial, the same reassignment was performed for all neurons. (A) Projections of the neural population state for each data set after a single shuffle trial. (B) The fraction of variance captured by the rotational model for the original data sets and 300 shuffle trials is shown (each shuffle trial tests a different random set of pairings of pre-movement and movement epoch activity). Error bars represent the standard deviation across shuffle trials. As with the second shuffle control, this manipulation is not expected to remove all rotational structure. While the relationship between rotational phase and the initial state is disrupted (as each condition's movement epoch activity is assigned to another condition's pre-movement activity), there is some likelihood that a given condition will be assigned an initial state similar to its own, allowing some rotational structure to be preserved by this manipulation. DOI:
Figure 3.
Figure 3.. Consistency of the rotational dynamics across conditions.
Traces represent histograms of the angle, q, between the neural state, x, and its derivative, x˙, for each time step. The angle was measured as illustrated schematically (inset) after projecting the data into the first jPCA plane. Purely rotatory activity results in angles near pi/2, while pure scaling/expansion results in angles near 0 or pi. Y-axis denotes scale for participant data (colored traces). For comparison, histograms for the example shuffle control data (Figure 2) and monkey composite data are shown in gray and black, respectively. These traces are normalized to match the participant data range (monkey data reproduced from Churchland et al., 2012). DOI:

Similar articles

See all similar articles

Cited by 16 PubMed Central articles

See all "Cited by" articles


    1. Churchland MM, Cunningham JP, Kaufman MT, Ryu SI, Shenoy KV. Cortical preparatory activity: representation of movement or first cog in a dynamical machine? Neuron. 2010;68:387–400. doi: 10.1016/j.neuron.2010.09.015. - DOI - PMC - PubMed
    1. Churchland MM, Cunningham JP, Kaufman MT, Foster JD, Nuyujukian P, Ryu SI, Shenoy KV. Neural population dynamics during reaching. Nature. 2012;487:51–56. doi: 10.1038/nature11129. - DOI - PMC - PubMed
    1. Churchland MM, Shenoy KV. Temporal complexity and heterogeneity of single-neuron activity in premotor and motor cortex. Journal of Neurophysiology. 2007;97:4235–4257. doi: 10.1152/jn.00095.2007. - DOI - PubMed
    1. Fetz EE. Are movement parameters recognizably coded in the activity of single neurons? The Behavioral and Brain Sciences. 1992;15:679–690.
    1. Georgopoulos AP, Kalaska JF, Caminiti R, Massey JT. On the relations between the direction of two-dimensional arm movements and cell discharge in primate motor cortex. The Journal of Neuroscience. 1982;2:1527–1537. - PMC - PubMed

Publication types

Associated data