Discovery of Brainwide Neural-Behavioral Maps via Multiscale Unsupervised Structure Learning

Vogelstein, Joshua T.; Park, Youngser; Ohyama, Tatsuya; Kerr, Rex; Truman, James W.; Priebe, Carey E.; Zlatic, Marta

doi:10.1126/science.1250298

Cited by 235 publications

(241 citation statements)

References 34 publications

Supporting

Mentioning

237

Contrasting

Order By: Relevance

“…Surprisingly there was little difference between the repertoires of wild type and proprioceptive mutant flies. Future work building on these results will include: (1) comparison of the methods we considered here with altogether different approaches, such as structure learning (Vogelstein et al 2014), (2) characterization of the generality of these findings Figure 9. Suggested unsupervised behavioral mapping method flow chart-the right branch represents an option for computing fast clustering by PCA-GMM-SW using a value of k informed by a first round of exploratory analysis using tSNE 2 -watershed.…”

Section: Resultsmentioning

confidence: 99%

“…Berman et al reported the first unsupervised mapping of adult Drosophila behavior from video data using probability density estimation to identify modes in time-frequency transformed data, thereby identifying stereotyped postural dynamics. Vogelstein et al (2014) used unsupervised structure learning to infer a hierarchical organization of larval behaviors based on eight time varying measures of posture and motion. Most recently an 'eigenlarva' analysis revealed that many behavioral events defy easy assignment to discrete clusters, suggesting that, at least in larvae, behavior may vary rather continuously (Szigeti et al 2015).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Systematic exploration of unsupervised methods for mapping behavior

Todd

Kain²,

Bivort

2017

Phys. Biol.

View full text Add to dashboard Cite

To fully understand the mechanisms giving rise to behavior, we need to be able to precisely measure it. When coupled with large behavioral data sets, unsupervised clustering methods offer the potential of unbiased mapping of behavioral spaces. However, unsupervised techniques to map behavioral spaces are in their infancy, and there have been few systematic considerations of all the methodological options. We compared the performance of seven distinct mapping methods in clustering a wavelettransformed data set consisting of the x-and y-positions of the six legs of individual flies. Legs were automatically tracked by small pieces of fluorescent dye, while the fly was tethered and walking on an air-suspended ball. We find that there is considerable variation in the performance of these mapping methods, and that better performance is attained when clustering is done in higher dimensional spaces (which are otherwise less preferable because they are hard to visualize). High dimensionality means that some algorithms, including the non-parametric watershed cluster assignment algorithm, cannot be used. We developed an alternative watershed algorithm which can be used in high-dimensional spaces when a probability density estimate can be computed directly. With these tools in hand, we examined the behavioral space of fly leg postural dynamics and locomotion. We find a striking division of behavior into modes involving the fore legs and modes involving the hind legs, with few direct transitions between them. By computing behavioral clusters using the data from all flies simultaneously, we show that this division appears to be common to all flies. We also identify individual-to-individual differences in behavior and behavioral transitions. Lastly, we suggest a computational pipeline that can achieve satisfactory levels of performance without the taxing computational demands of a systematic combinatorial approach.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Systematic exploration of unsupervised methods for mapping behavior

Todd

Kain²,

Bivort

2017

Phys. Biol.

View full text Add to dashboard Cite

show abstract

“…To illustrate, consider the project of Vogelstein et al 7 , whose aim was to understand in Drosophila larvae the causal role of each of 10,000 neurons in producing a simple behavior in the animal's repertoire, such as turning or going backwards. Drawing on over 1,000 genetic lines and using optogenetic techniques to stimulate individual neurons in each line, they generated a basic data set consisting of correlations between stimulated identified neurons and a behavioral output.…”

mentioning

confidence: 99%

Putting big data to good use in neuroscience

Sejnowski¹,

Churchland²,

Movshon³

2014

Nat Neurosci

268

181

View full text Add to dashboard Cite

Big data has transformed fields such as physics and genomics. Neuroscience is set to collect its own big data sets, but to exploit its full potential, there need to be ways to standardize, integrate and synthesize diverse types of data from different levels of analysis and across species. This will require a cultural shift in sharing data across labs, as well as to a central role for theorists in neuroscience research.Big data, the buzz phrase of our time, has arrived on the neuroscientific scene, as it has already in physics, astronomy and genomics. It offers enlightenment and new depths of understanding, but it can also be a bane if it obscures, obstructs and overwhelms. The arrival of big data also marks a cultural transition in neuroscience, from many isolated 'vertical' efforts applying single techniques to single problems in single species to more 'horizontal' efforts that integrate data collected using a wide range of techniques, problems and species. We face five main issues in making big data work for us.First, data in neuroscience exist at an astonishing range of scales of both space and time. Neuroscientific data are obtained from a wide range of techniques, from patch clamping to optogenetics to fMRI (Fig. 1). Most of these techniques are used one at a time. One lab will record spikes from an array of neurons, but not be able to determine which types of neurons they are or how they are connected to other neurons. Another lab will reconstruct the wiring diagram of the same circuit, but without recording data to identify the properties of the reconstructed neurons. In some heroic cases, functional data have been laboriously combined with anatomical reconstructions 1 , but rarely if ever in a broad behavioral context.

show abstract

“…Computational models applied to such data will facilitate both interpreting population neural activity and connecting neural activity with behavior. In parallel, the use of unsupervised classification methods has revealed stereotyped structure in animal behavior -an animal's movements over time can be described as sequences of discrete behavioral modules (Vogelstein et al 2014;Berman et al 2014;Wiltschko et al 2015). Computational modeling's task will now be to determine how sensory cues and internal states affect behavioral sequencing, and how neural codes underlie the choice of behavioral modules.…”

Section: Prospectsmentioning

confidence: 99%

The use of computational modeling to link sensory processing with behavior in Drosophila

Clemens

Murthy

2017

Preprint

View full text Add to dashboard Cite

A major goal of systems neuroscience is to understand how the brain represents information, and how those representations are used to drive behavior. Even though genetic model organisms like Drosophila grant unprecedented experimental access to the nervous system for manipulating and recording neural activity, the complexity of natural stimuli and natural behaviors still poses a challenge for solving the connections between neural activity and behavior. Here, we advocate for the use of computational modeling to complement (and enhance) the Drosophila toolkit. We first lay out a modelling framework for making sense of the relation between natural sensory stimuli, neuronal responses, and natural behavior. We then highlight how this framework can be used to reveal how neural circuits drive behavior, using selected case studies.PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2720v1 | CC BY 4.0 Open Access |

show abstract

Discovery of Brainwide Neural-Behavioral Maps via Multiscale Unsupervised Structure Learning

Cited by 235 publications

References 34 publications

Systematic exploration of unsupervised methods for mapping behavior

Systematic exploration of unsupervised methods for mapping behavior

Putting big data to good use in neuroscience

The use of computational modeling to link sensory processing with behavior in Drosophila

Contact Info

Product

Resources

About