UCLA Miniscope - User contributions [en]

GRIN Lens Information

2018-08-28T13:48:33Z

Thebesteagle: /* GRIN Lens Issues */

[[File:GRINProfil.png|thumb|300px|Example index of refraction profile of a 2mm diameter lens.]]
GRadient INdex (GRIN) lenses use a non uniform index of refraction to shape light instead of surface curvature as with standard lenses. Our system uses a cylindrical GRIN lens which is implanted within 200um of the structure of interest. For a given GRIN lens, its focusing property can be adjusted by changing the pitch of the lens which is directly related to the lens length.
*Advantages of GRIN lenses
**Extremely small to zero working distance
**Optical properties tune-able by changing lens length
**Affordable
**Small size (down to 0.5mm in diameter)
**Compatible
*Disadvantages of GRIN lenses
**No optimal index of refraction profile for imaging applications
**The majority of commercially available GRIN lenses are not optimized for imaging applications (discussed below)
**People have experienced supply issues from some suppliers (discussed below)

Our system currently uses a 0.25 pitch GRIN lens along with an achromatic lens to form an imaging on the microscope's imaging sensor. Sliding the imaging sensor up or down will shift the focal plane in the brain. We have successfully imaged hippocampus CA1, visual cortex, and subiculum using 2mm and 1.8mm diameter GRIN lenses. We are now in the process of testing smaller diameter relay lenses which will work in conjunction with the larger 0.25 pitch lens.
== GRIN Lens Optics ==
A nice overview of GRIN optics can be found on [[http://www.gofoton.com/product/selfoc-imaging-lenses/]] (click on "Physics of SELFOC") and [[http://www.grintech.de/gradient-index-optics.html|Grintech.de]]. The most common use for GRIN lenses is for fiber coupling and light collimation and focusing. Aside from Grintech, all other GRIN lens manufacturers appear to produce GRIN lenses mainly for this use and not specifically for imaging. The index of refraction profile for most commercial GRIN lenses follow a second order expansion of a hyperbolic secant curve. The second order expansion (a parabola) is the ideal profile for on axis focusing of collimated light but has focusing issues of off axis light. Grintech uses an additional manufacturing step to add the fourth order expansion term of the hyperbolic secant to their index of refraction profile. The fourth order term improves off axis focusing which is very important for imaging applications. Theoretically there is not an ideal index of refraction profile for an imaging GRIN lens but the 4th order expansion term does a good enough job for imaging in our system.

Information about GRIN lens optics and manufacturing is a bit hard to find. We have listed a few papers below which provide nice insight on the topic
*Analysis of Refractive Index Distributions in Cylindrical, Graded- Index Glass Rods (GRIN Rods) Used as Image Relays
**E. G. Rawson, D. R. Herriott, and J. McKenna
**March 1970 / Vol. 9, No. 3 / APPLIED OPTICS
*Diffraction-limited gradient-index (GRIN) microlenses with high numerical apertures produced by silver ion exchange in glass: diffusion modeling and process optimization
**Bernhard Messerschmidt, Ulf Possner, Albrecht v. Pfeil, and Torsten Possner
**SPIE Vol. 3424 s 0277-786X/98
*Fabrication of gradient refractive index rod lens using double ion exchange processes
**Hao Lv, Aimei Liu, Jufang Tong, Xunong Yi, Qianguang Li, Xinmin Wang, Yaoming Ding
**Optical Engineering 50(7), 073402 (July 2011)
*

== GRIN Lens Issues ==
As mentioned in the section above, the majority of commercially available GRIN lenses are manufactured with a second order sech profile causing poor off-axis focusing. While these lenses can resolve similar sized objects as fourth order sech profile GRIN lenses, they have much lower contrast. This is due to some angles of a point source of light not being focused to the same 3-dimensional spot as other angles of light from that same point source. This results in a nicely resolved image generated by some angles of light with a diffuse glow of light generated by other light angles. This proves to be a reasonably big issue when imaging neural activity since the unfocused, diffuse glow produced by most GRIN lenses increases your measured background fluorescence, making it difficult to pull out dF/F activity.

Currently there seems to be only one manufacturer, Grintech, that produces GRIN lenses with a 4th order index of refraction profile. Additionally, Grintech uses an ion exchange process that has been certified to be bio-compatible. That being said, we have heard from many people that it can be difficult to get lenses from Grintech due to production limitations or exclusive sales agreements. For that reason we are actively working with other GRIN lens manufacturers to modify their product lines to produce high quality imaging GRIN lenses. More information on this topic can be found in the section below.

== GRIN Lens Manufacturers/Suppliers==
There are three major manufacturers of GRIN lenses world wide.
;Grintech
:Currently Grintech produces the best lenses for imaging applications but many people have experienced supply issues with them. The work presented on this site has all been done using Grintech lenses.
;Go!Foton (NSG)
:Likely the largest manufacturer of GRIN lenses. Their product lines are mainly aimed at telecommunication applications and fiber coupling. They do sell an 'imaging' version of their lens and we are actively testing out these new lenses. We are also working with them to modify their production lines to produce high quality imaging GRIN lenses.
;Altechna/Chinese manufacturing
:Altechna is a Lithuanian company with manufacturing in China. We have been in talks with them to improve the imaging properties of their lenses.

== GRIN Lens Specifications for the Miniscope System ==
*Our system currently uses 2mm or 1.8mm diameter GRIN lenses with a pitch of (~0.25 + 0.5*N), where N is an integer.
*We have successfully used 1.8mm and 2mm GRIN lenses from Grintech (which are about 4mm to 5mm in length) to image hippocampal CA1. Their respective part numbers are GT-IFRL-180_inf_50-NC and GT-IFRL-200_inf_50-NC.
**If you are unsuccessful ordering these lenses from Grintech you can obtain 1.8mm diameter rebranded Grintech lenses from [http://www.edmundoptics.com/optics/optical-lenses/aspheric-lenses/gradient-index-grin-rod-lenses/3145/#f=categories_s|*C11I*,productId_i|3145,27614_s|1.8 Edmund Optics]. We have successfully tested part #64-519 but most 1.8mm diameter GRIN lenses sold by Edmund Optics should work.
***You want to pick a lens that has a length between ~3.93mm and 4.31mm.
***We have not tested any of the coated lenses.
*GRIN lenses with a 4th order index of refraction profile will image significantly better than 2nd order lenses.
*The overall length of your GRIN lens should stick far enough out of the skull to allow you to hold it during implantation and cement it to the skull.
*The distance between the top of the GRIN lens and other miniscope optics is not too critical. You will want to cement the baseplate so that the top of the GRIN lens is somewhere between the bottom of the scope body and the internal dichroic mirror.
*For GRIN lenses smaller than 1.8mm in diameter, a GRIN relay lens should be used as the implanted lens and then imaged with a 2mm diameter GRIN lens. We are currently working on this.

== Guide to using 1mm and smaller GRIN lenses ==
Due to the optical properties of GRIN lenses, it is only feasible to use a single, 0.25pitch, GRIN lens for larger (1.8mm or greater) diameter lenses. To image with thinner (1mm or smaller) GRIN lenses, a simple modification must be made to the scope body to allow for a dual GRIN lens setup. [[Imaging With Thin GRIN Lenses]] will walk your through all aspects of this approach.

Head Mounted Scope

2018-08-17T12:50:56Z

Thebesteagle: /* Overview */

[[File:OverviewScope.png|thumb||600px]]

== Overview ==
[[File:MiniscopeExplodedLabled.PNG|thumb|right|300px|Exploded view of the head mounted scope.]]
The head mounted scope consists of a machined Delrin housing, optical filters and lenses, an excitation light source, and CMOS imaging sensor. The mass of the system is under 3 grams and connected to the DAQ hardware using a single coaxial cable. The body and optical components can be easily hand assembled using only forceps and a torx T2 driver. Optical filters are slid in from the side of the scope and held in place with a filter cover plate. The excitation half sphere lens sits in a spherical cutout and held in place by pressure from the LED PCB (optical glue is optional). The achromatic lens is slid through the top emission hole and either press fit into place or optical glued.

Adjustment of focal plane is achieved through adjusting the distance between the CMOS imaging sensor and other optical elements. The CMOS imaging sensor is mounted onto a sliding focusing mechanism which is moved by hand and locked in place with a 00-80 setscrew. Roughly speaking, movement of about 5mm of the CMOS imaging sensor results in a focal plane change of ~150um.

[[File:MiniscopeCalibrationSlide.png|thumb|left|300px|Image of a calibration slide with 9.8μm line spacing (superimposed red boxes are 10px x 10px)]]
The scope attaches to the head of the animal with the use of a small aluminum baseplate. The baseplate and scope have a matching set of 3 rare-earth magnets which helps with mounting the scope on an awake animal. Once the scope is in place a 00-80 set screw is used to fix it in place.

 

== Optical Path ==
[[File:MiniscopeOpticalPath.png|thumb|300px|Cross section of scope. Excitation path is in blue. Emission path is in green.]]
The optical path of the miniature microscope is very similar to a tabletop wide-field fluorescence microscope. The main difference is the size of the optical elements and the use of a GRIN lens as an objective. A more detailed discussion on GRIN lenses can be found [[GRIN Lens Information|here]].

The optics in the miniscope use off the shelf lenses and diced commercial filters. Excitation light generated by an LED passes through a half-ball lens to help collimate the light
which is then bandpassed through an excitation filter and reflected using a dichroic mirror.

Emission light exits the objective GRIN lens in roughly a parallel orientation, passes through the dichroic mirror and then bandpassed through an emission filter. The light is then focused onto the CMOS imaging sensor using an achromatic lens.

 

== Machined Delrin Housing ==
The housing of the miniature microscope is machined out of plastic (we suggest using Delrin) and consists of a main body, filter set holder, and focusing slider.
 

== Aluminum Baseplate ==
The baseplate is cemented to the animal's skull and provides the interface for temporarily connecting the miniature microscope to the animal. Three pairs of magnets on the baseplate and bottom of the microscope help snap the microscope in place and aid in placement when working with awake, freely moving animals. Once the microscope is placed in the baseplate, a 00-80 setscrew locks the scope in place. The setscrew applies a force along both sides of the scope making the mount very stable.

 

== CMOS Imaging Sensor PCB ==
[[File:CMOSPCB_1.png|thumb|300px|PCB layout of the CMOS imaging sensor PCB. The PCB is 13mm wide and should be printed on a 0.031" circuit board.]]
The CMOS imaging sensor PCB holds all the electronics needed to power and control the CMOS imaging sensor and excitation LED. A list of circuit features and components are listed below
*Power over coax filter to separate data/control stream from DC power.
*3.3V and 1.8V power regulators.
*Red LED shows board is powered properly and can be used for animal tracking.
*Serializer interfaces with imaging sensor and packages data to be sent over coax.
*A Digital to Analog Converter (DAC) and constant current source are used to control and power the excitation LED.

[[File:CMOSSchematic.png|thumb|left|500px|Schematic of CMOS imaging sensor PCB.]]
 

== Coaxial Cabling ==
[[File:CoaxFPDLink.png|thumb|300px]]
The head mounted scope is connected to the DAQ hardware through a single, lightweight coaxial (coax) cable.

'''50Ω Coaxial Cable Features'''
*Power over coax + control signals + data + GPO
*Down to 0.3mm diameter
*Up to 15m long
*Commutator compatible
*Addition 4 lines of GPO
*Supports 12bit pixel data
*Up to 1.4Gbps

== Excitation LED PCB ==
The excitation LED PCB is a simple 2 sided PCB with solder pads for a Luxeon SMD LED (P/N LXML-PB01-0030) and wires to power the LED. The PCB acts as a small heat sink and has a 1mm mounting hold for attaching it to the miniscope body. The LED is powered by a constant current source on the CMOS imaging sensor PCB and controlled through an 8bit DAC. We generally run this LED at only a few percent of it maximum power (3mA to 10mA current).
[[File:LEDPCB.png|thumb|300px|Top and bottom of the LED PCB. The PCB is 7mm tall and should be printed on 0.031" thick circuit boards.]]
 

== Mass of Head Mounted Scope Components ==
{| class="wikitable"
|-
! scope="col"| Component
! scope="col"| Mass (g)
! scope="col"| Quantity per Scope
! scope="col"| Total Mass
! scope="col"| Notes
|-
! Main Body
| 0.51
| 1
| 0.51
|
|-
! Focusing Slider
| 0.48
| 0.8 to 1
| 0.38 to 0.28
| Bottom portion can be cut off to reduce weight.
|-
! Filter Cover
| 0.11
| 1
| 0.11
|
|-
! Screw (3mm long 1mm diameter)
| 0.025
| 3 to 7
| 0.075 to 0.175
|
|-
! Magnet
| 0.012
| 3
| 0.036
|
|-
! Optical Filter Set (per 1mm x 1mm x 1mm of glass)
| 0.0017
| 54
| 0.09
|
|-
! Achromatic Lens (5mm Diameter)
| 0.15
| 1
| 0.15
| Varies slightly with different focal lengths
|-
! Half Ball Lens
| 0.02
| 1
| 0.02
|
|-
! LED PCB (without LED SMD)
| 0.07
| 1
| 0.07
|
|-
! LED SMD + Wires
| 0.05
| 1
| 0.05
|
|-
! CMOS PCB (excluding CMOS sensor)
| 0.85
| 1
| 0.85
|
|-
! CMOS Imaging Sensor
| 0.76
| 1
| 0.76
|
|}

*Total scope mass between 3g and 3.3g.
*Around 0.6g of the CMOS PCB comes from the circuitry and components needed to support power and data over a single coaxial cable. In our experience a very light, flexible cable is well worth the trade off of having the scope be 0.6g lighter.
*It is feasible to drop the scope's mass to under 3g with a little bit of work. It could be dropped closer to 2g if the bare die of CMOS imaging sensor was directly wire bonded to the PCB.

FAQs

2018-08-17T08:26:02Z

Thebesteagle: /* Where can I find more information on GRIN lenses and GRIN lens related issues */

Below you will find answers to common questions about our microscope system. Is you have a question that is not address here consider asking it on our [[Special:Wikiforum|Discussion Board]].

== Are you selling your system? ==
We are not selling any part of our system. Rather, we have provided the design files, specifications, and part numbers so that anyone can build a system. While we understand many people may prefer an off-the-shelf product, we have made every effort to make our system as easy to build as possible.

== How much does your system cost? ==
In general, our miniscope system will cost around $3-5K for 5 miniscopes, a DAQ board, and all associated hardware and software. Each additional scope will cost ~$400. You can find a general breakdown of costs on the master parts list. In general, costs can be substantially reduced by buying in bulk so we highly recommend partnering with other labs to reduce costs.

== What skill sets are needed to setup a miniscope system of my own? ==
Basic soldering knowledge is needed for soldering wires to PCBs as well as possibly soldering connectors to coax cables. Some great soldering tutorials can be found [https://www.sparkfun.com/tutorials/category/2 here]. Physically assembling the miniscope system does not require any specialized knowledge or skills.

== What support do you offer if we have trouble with our miniscope system? ==
This wiki is designed to answer the most common questions that arise from using our miniscope system. We have a [[Special:WikiForum|Discussion Board]] where any user can post and answer any questions about any aspect of the system and we will monitor this board as much as possible. While we do not have the resources to personally troubleshoot every problem, we hope that the collective knowledge from the community of users will allow everyone to solve any problems they encounter.

== What design software do you suggest using to modify the miniscope system? ==
The machined plastic parts of the scope are designed using SolidWorks which is a powerful, easy to use and affordable 3D CAD program.

[http://www.cypress.com/documentation/software-and-drivers/ez-usb-fx3-software-development-kit Cypress EZ-USB FX3 SDK] contains the programs needed to modify the DAQ firmware, GPIF II interface, and flash the firmware to the DAQ hardware.

I use Microsoft Visual Studio to develop our DAQ software which can be downloaded for free through [https://www.dreamspark.com/ Microsoft DreamSpark] but is not necessary for writing your own DAQ software. The miniscope hardware enumerates as a generic webcam which means it can be controlled through other open source and commercial webcam software. OpenCV libraries also provide a nice starting point for building your own DAQ software.

== Where can I find all the design files related to the miniscope project? ==
All downloadable content can be found on our [[Files for Download]] page.

== Where can I find more information on GRIN lenses and GRIN lens related issues? ==
Take a look at [[GRIN Lens Information]].

== How does the Miniscope system differ from the nVista system from Inscopix? ==

;Advantages of our Miniscope system:
:'''Cabling:''' The Miniscope system using a single coaxial cable to connect the scopes to the DAQ box. These coax cables are commercially available, very robust, cheap, lightweight, and flexible.
:'''Commutator compatible:''' Due to our cabling and data protocol used in our system, a commutator can be hooked up between the scope and the DAQ box minimizing cable and animal strain and expanding the possibilities of what can be done with miniature microscopes. We have successfully used a 2 channel commutator from DragonFly.
:'''Open-source:''' Our system is the opposite of a black-box. We think science benefits the most when scientists understand what their equipment/analysis is doing and how they can modify it to best suit their needs.
:'''Cost:''' Building a Miniscope system consisting of 4 scopes and 1 DAQ box will cost around $3k. A comparable system from Inscopix costs around $250k. Most labs can afford to build a Miniscope system for each researcher interested in using it. This removes the complications that arise from sharing equipment during time sensitive experiments.
:'''Focusing Mechanism:''' The Miniscope system uses a sliding, rather than rotating, focusing mechanism for adjusting the imaging plane. No rotational imaging registration is needed and we have noticed that finding the correct imaging plane is easier with a sliding focus.

;Advantages of the nVista system from Inscopix:
:Works out of the box. No need for assembly.
:'''Support:''' Inscopix provides commercial support for their system.

== Can the Miniscope body be 3D printed? ==
Yes but we do not recommend it. CNC machined Delrin plastic can achieve greater strength and precision, thinner walls, and is generally cheaper than 3D printed parts. Also, Delrin is chemically resistant so you can clean them with acetone.

FAQs

2018-08-17T08:22:25Z

Thebesteagle: /* What design software do you suggest using to modify the miniscope system? */

Below you will find answers to common questions about our microscope system. Is you have a question that is not address here consider asking it on our [[Special:Wikiforum|Discussion Board]].

== Are you selling your system? ==
We are not selling any part of our system. Rather, we have provided the design files, specifications, and part numbers so that anyone can build a system. While we understand many people may prefer an off-the-shelf product, we have made every effort to make our system as easy to build as possible.

== How much does your system cost? ==
In general, our miniscope system will cost around $3-5K for 5 miniscopes, a DAQ board, and all associated hardware and software. Each additional scope will cost ~$400. You can find a general breakdown of costs on the master parts list. In general, costs can be substantially reduced by buying in bulk so we highly recommend partnering with other labs to reduce costs.

== What skill sets are needed to setup a miniscope system of my own? ==
Basic soldering knowledge is needed for soldering wires to PCBs as well as possibly soldering connectors to coax cables. Some great soldering tutorials can be found [https://www.sparkfun.com/tutorials/category/2 here]. Physically assembling the miniscope system does not require any specialized knowledge or skills.

== What support do you offer if we have trouble with our miniscope system? ==
This wiki is designed to answer the most common questions that arise from using our miniscope system. We have a [[Special:WikiForum|Discussion Board]] where any user can post and answer any questions about any aspect of the system and we will monitor this board as much as possible. While we do not have the resources to personally troubleshoot every problem, we hope that the collective knowledge from the community of users will allow everyone to solve any problems they encounter.

== What design software do you suggest using to modify the miniscope system? ==
The machined plastic parts of the scope are designed using SolidWorks which is a powerful, easy to use and affordable 3D CAD program.

[http://www.cypress.com/documentation/software-and-drivers/ez-usb-fx3-software-development-kit Cypress EZ-USB FX3 SDK] contains the programs needed to modify the DAQ firmware, GPIF II interface, and flash the firmware to the DAQ hardware.

I use Microsoft Visual Studio to develop our DAQ software which can be downloaded for free through [https://www.dreamspark.com/ Microsoft DreamSpark] but is not necessary for writing your own DAQ software. The miniscope hardware enumerates as a generic webcam which means it can be controlled through other open source and commercial webcam software. OpenCV libraries also provide a nice starting point for building your own DAQ software.

== Where can I find all the design files related to the miniscope project? ==
All downloadable content can be found on our [[Files for Download]] page.

== Where can I find more information on GRIN lenses and GRIN lens related issues ==
Take a look at [[GRIN Lens Information]].

== How does the Miniscope system differ from the nVista system from Inscopix? ==

;Advantages of our Miniscope system:
:'''Cabling:''' The Miniscope system using a single coaxial cable to connect the scopes to the DAQ box. These coax cables are commercially available, very robust, cheap, lightweight, and flexible.
:'''Commutator compatible:''' Due to our cabling and data protocol used in our system, a commutator can be hooked up between the scope and the DAQ box minimizing cable and animal strain and expanding the possibilities of what can be done with miniature microscopes. We have successfully used a 2 channel commutator from DragonFly.
:'''Open-source:''' Our system is the opposite of a black-box. We think science benefits the most when scientists understand what their equipment/analysis is doing and how they can modify it to best suit their needs.
:'''Cost:''' Building a Miniscope system consisting of 4 scopes and 1 DAQ box will cost around $3k. A comparable system from Inscopix costs around $250k. Most labs can afford to build a Miniscope system for each researcher interested in using it. This removes the complications that arise from sharing equipment during time sensitive experiments.
:'''Focusing Mechanism:''' The Miniscope system uses a sliding, rather than rotating, focusing mechanism for adjusting the imaging plane. No rotational imaging registration is needed and we have noticed that finding the correct imaging plane is easier with a sliding focus.

;Advantages of the nVista system from Inscopix:
:Works out of the box. No need for assembly.
:'''Support:''' Inscopix provides commercial support for their system.

== Can the Miniscope body be 3D printed? ==
Yes but we do not recommend it. CNC machined Delrin plastic can achieve greater strength and precision, thinner walls, and is generally cheaper than 3D printed parts. Also, Delrin is chemically resistant so you can clean them with acetone.