Long-term dynamics of CA1 hippocampal place codes Suzy Xu and - - PowerPoint PPT Presentation

Long-term dynamics of CA1 hippocampal place codes

Suzy Xu and Emika Lisberger BioNB 4110 April 21, 2014

Nature Neuroscience

• Published in Volume 16,

Issue 3

• Subset of Nature

Publishing Group

• Impact Factor (as of 2012)

is 15.251

• Ranked 6th out of 251

journals in the “Neuroscience” category

• Dr. Mark Schnitzer (PI)
• Education:
• BA in physics from Harvard
• MA in physics from Princeton
• PhD in physics from Princeton
• Positions:
• Investigator of the Howard Hughes Medical Institute
• Associate Professor of Biology and Applied Physics at Stanford

University

• Current Research:
• In vivo two-photon fluorescence imaging studies of cerebellar-

dependent learning and memory

• Fiber optic fluorescence microendoscopy of the hippocampus,

thalamus, and inner ear

• Massively parallel brain imaging in live fruit flies

Designed experiments, wrote the paper, and supervised the study

Yaniv Ziv

• Education:
• B.S. in Biology from The Hebrew University of

Jerusalem in 2001

• PhD in Neurobiology from the Weizmann

Institute of Science in 2007

• Position:
• Postdoc under Mark Schnizter in the

Department of Biology at Stanford University

• Current Research:
• Effects of experience on structural and

functional interactions between different types of hippocampal cells in vivo

Designed experiments, acquired data, and wrote the paper

Laurie D. Burns

• Education:
• B.S. in Physics from MIT
• PhD in Applied Physics from

Stanford University in 2012, working

• n the development of miniature

microscope technology

• Current Work:
• Consultant at Inscopix Inc.
• Was previously one of the founding

scientists

Designed experiments, acquired data, analyzed data, built equipment and wrote the paper

Eric D. Cocker

• Education:
• B.S., M.S., and PhD in

Mechanical Engineering at Stanford University

• Current Work:
• A member of the Founding

team at Inscopix, as the Founding Principal Engineer

Built the equipment used

Elizabeth O. Hamel

• Acquired the data used in this study.
• No other information found online

Kunal K. Ghosh

• Education:
• B.S. at The Wharton School

(University of Pennsylvania)

• B.S.E in Electrical Engineering at

University of Pennsylvania

• M.S. and PhD in Electrical

Engineering at Stanford University

• Postdoc in the Department of

Biology at Stanford

• Current Work:
• Founder and CEO of Inscopix Inc.

Built the equipment used

Lacey J. Kitch

• Education:
• B.S. in Physics and Math with

Computer Science at MIT

• M.S. in Electrical Engineering at

Stanford University

• Current Work:
• Pursuing a PhD in Electrical

Engineering at Stanford

• Interested in neural computation
• f processes in living brains.

Analyzed and wrote the paper

Abbas El Gamal

• Education:
• B.S. Honors in Electrical Engineering at Cairo

University

• M.S. in Statistics and Ph.D. in Electrical

Engineering at Stanford University

• Current Position:
• Hitachi America Professor in the School of

Engineering

• A member of the National Academy of Engineering

and a Fellow of the IEEE (Institute of Electrical and Electronics Engineers)

• Plays key roles in several Silicon Valley companies.
• Research Contributions:
• Include information theory, Field Programmable

Gate Array and digital imaging devices and systems

Supervised the study

The Experiment

• Abstract: “Using Ca2+ imaging in freely behaving

mice that repeatedly explored a familiar environment, we tracked thousands of CA1 pyramidal cells’ place fields.”

• Goal: To find out the long-term stability of place

fields using one-photon imaging

General Methods

• Used GCaMP3 to express Ca2+ in pyramidal cells by

injection of a viral vector into CA1

• Used miniaturized microendoscope for Ca2+ imaging in

four freely behaving mice

• Tracked Ca2+ dynamics of 515 to 1,040 pyramidal cells

per mouse on repeated visits to a familiar track

• Used water rewards to train the mice to run up and down

the track

• Recorded for 45 days

One-Photon Microendoscopy

• Imaging technique used in this study
• Inserted optical fiber that acts as a

lens into brain tissue

• Lens diffracts light to one point
• Base stayed attached on the mouse

brain for 45 days

Two-Photon Microscopy

• Basics:
• Developed by Winfried Denk in the lab of Watt W

. Webb at Cornell University in 1990

• Allows imaging of live tissue up to 1.6 mm in

depth

• Two Photon vs. One Photon:
• 2 photons of half the energy are excited

simultaneously

• More localized excitation
• Fluorescence photon is emitted
• Benefits:
• Images only what is labeled with a fluorescent dye
• Less energy less damage to sample
• Longer wavelength less scattering (better

resolution along z-axis)

• The Schnitzer lab is working to incorporate two-

photon microscopy into their microendoscopes

E = hv = hc λ

I ∝ 1 λ 4

Three-Photon Microscopy

• Developed by Xu lab at Cornell
• Can image individual neurons
• Can image hippocampus without removing overlying

tissue

• Figure of biological imaging!

Hippocampus (Hp)

• Under the cerebral cortex, and

in the medial temporal lobe

• Information travel in the tri-

synaptic pathway

• Responsible for episodic

memory and context processing

• Intact Hp is especially

important for responding to information about spatial relations

Place Fields

Place cells in an intact hippocampus form place fields when an animal is put into a novel environment

Initial Imaging: Calcium activity

Consistent over time

• No damage to cells
• Microscope is accurate

Distribution of size and location

Direction Preference

• Data pooled from four mice,
• n day 15
• Rearranged firing data for this

graph

• Left place cells fire for leftward

movement only (c)

• Right place cell fire for

rightward movement only (d)

Which cells fire?

• Pooled data from 4 mice
• Gaussian smoothed

density of overlapping right and left movement

• Place fields were

consistent throughout the whole experiment, but rearranged

• 20% of cells that are place

cells for leftward and rightward movement

Decrease of active cells

• Ca2+ activity of 826 cells

in one mouse over 45 days (a)

• The number of sessions a

cell is active for (b)

• Probability of recurrence

from session to session of place fields declines with time (c)

• When the place fields are

recurrent, their locations are generally identical (d)

15-25% Recurrence

After the experiment

Why is there recurrence?

• Determined that it was not because of physiological or coding

parameters (Ca2+ activity patterns)

Is the 15-25% overall recurrence sufficient to retain a stable spatial representation?

• Using Bayesian decoding, they determined if they could

reconstruct the mouse’s location from the Ca2+ imaging data

Bayesian Decoding

• What is it?
• Using the arrival times of

Ca2+ activity and the probability of seeing a certain stimulus to develop a neural code

Bayesian Decoding

• Figure h
• Same-day decoding used

data from the same day to decode place

• Time-lapse decoding used

data from day 5 to decode place in days 10, 20, and 35

• Figure i
• Shows median error over time
• Same day: ~8% error
• Time lapse: ~15% error
• Figure j
• Cumulative percentage of error

Potential Flaws

• Data
• The authors disregarded the fact that GCaMP3 does

not record single spikes, but instead bursts of spikes.

• Analysis
• How do we know from this data that you can get a

stable spatial recognition?

How could this experiment be improved?

Discussion Questions

• What are place cells? Where are they found in the brain?
• What is the optical imaging method used in this study? Briefly

describe how it works.

• How did the authors test whether the 15-25% place cell

recurrence was sufficient to determine the mouse’s location?

• Why is this paper so progressive, what does this contribute/

mean to future neuroscience research?

• What are some potential flaws in this paper? What future

experiments could these neurobiologists do to improve the results?