Wireless Ephys Aquisition, Intan and WiFi

[Josh Wardell]

I’m working on a project where we are looking to transmit ephys data wirelessly, and considering using an off-the-shelf WiFi module connected to an Intan RHD2000 headstage. I’m wondering if anyone has attempted something similar and can offer any suggestions or advice.

There are a few commercial options out there that seem to be in the range of 64 channels for $40k, we were hoping for more channels or much lower cost. Wireless range only needs to be about one meter, and while a prototype would be battery powered we would hope to eventually inductively power it from a distance of a few inches so everything can be sealed on the head. The goal is to never need to interrupt the monkey or remove it from its cage.

There are microcontroller wifi modules available for less than $10 that support SPI to communicate with the Intan headstage. On the computer side, a plugin could be made for Open Ephys to receive the data stream from IP.

Does this seem like an unreasonable path? Does anyone see any problems with attempting this or have thoughts on alternatives?

Thanks.

Josh Wardell

MIT Dept. of Brain and Cognitive Sciences Electronics Shop

http://46shops.mit.edu/eshop

[Aarón Cuevas]

Hi Josh,

My experience says that a normal microcontroller has lots of trouble driving a RHD chip, specially trying to achieve very strict timing constraints, which is a must on data acquisition, unless you run it at slow sample rates (up to 1-2kS/s might not pose a lot of problems, but the challenge escalates quickly from there). Bulk DMA SPI transfers might mitigate this, but the RHD chip needs to deassert the CS line for each conversion and most microcontrollers either deassert it at the start of a bulk transfer or do not even drive that signal at all. Many regular multichannel ADCs work this way, though, so there probably are DSP with a DMA engine able to properly drive a RHD chip which could then transfer the data to the wifi module.

Another option would be to use a small low-power FPGA like Microsemi's igloo series or Lattice's iCE series. If you're planning to use a RHD2164 chip, with works with a DDR SPI data-transfer scheme (not quite standard) I'd chose this option.

Best,

Aarón.

[Robert Paugh]

Hi Josh,

We are developing a 32 channel + 3 aux channel WiFi Intan headstage.  Prototype is working in the lab.  35 channels of 20ksps data is continuously streaming via WiFi.  Streaming duration depends on battery size.  Pricing is not yet set but I do know it will be less than $40K!  

Inductive power options are being explored.  What would be your use case; streaming data 24hrs a day, or a one hour streaming experiment per day, or more like 10 minutes every hour, …. ?  

As Aarón mentions, with uC it is tricky to get a continuous CSn pulse with acceptable jitter.  We have solved this issue for our uC without needing an FPGA.

We also have other RF data links for the RHD2000 that have lower channel count (10).  These links are much lower power than WiFi.  

I'm located in the Boston area.  Happy to talk with you about your requirements.

Thank you for your time,

Robert 

[Josh Wardell]

Thanks Aaron, those are the unforeseen problems that I was worried about hitting halfway through development. I would start with just a few channels, and could add some more processing power if needed later, but clearly it won't be a quick simple project.

Robert, I would be very interested is speaking with you to go over more details. Power would only be needed for a few minutes once or twice a day, but the location will always be in the same spot so it should be relatively simple for inductive power. 32 channels might be enough if we could run more than one device at the same time, assuming wifi has the bandwidth to handle a few.

[Robert Paugh]

Hi Open-Ephys,

Jon Newman has asked that I clarify why the WiFi headstage that we are working on is useful to the open-Ephys community.

Our WiFi headstage HW and FW are closed source and will be for sale.  Price is still being worked out.

However, we are using a modified version of the open source Qt Intan GUI to interface with and control the headstage.  This modified Qt version will still be open source.  

We also plan to have an open source plugin for the open-Ephys GUI once we figure out how to make a plugin.

There is a simple API that will be provided so folks can make and maintain their own GUIs using C# or Octave, …

My invitation to Josh to go over requirements is open to all.  I welcome your input.  We are still working out features and options.

Sorry for the confusion,

Robert

[Robert Paugh]

Hi Josh,

Our WiFi-to-LVDS interface for Intan amplifier boards is heading to pilot assembly by the end of June.  The RCB-W24A-LVDS is a 0.85” x 1.75” module with 2.4Ghz 802.11n WiFi, SPI LVDS interface on 12-Pin Omnetics connector, and a JST 2mm battery connector.   We can stream up to 32 channels at 20ksps 16 bits continuously.  3 Aux channels and Vdd measurement are also included.

Connect your Intan amp board directly to the 12-Pin Omnetics or use a 1meter blue cable.  We have a modified version (subset) of the Intan Qt GUI that is working great with the RCB-LVDS module.  Qt GUI will be open source, same as Intan.  We will have an open API and hope to have an open-Ephys plugin.  

Price is $895.00 plus taxes and shipping for the module, includes a DC power cable.  Battery and enclosure are not included at this time.  For testing we use protected lithium bag batteries from Sparkfun or AdaFruit.  Streaming time depends on battery size.  Battery should be at least 150mAh.   

Some details and photos are on the website at www.dspwi.com.  More details and user manual available soon.

Thanks to you and Kohitij Kar for your help in testing and brainstorming features and behaviors.   And thanks to Reid Harrison at Intan for his help in getting us started.

Happy to discuss details with any open-Ephys member on the forum or ok to contact me directly for more information.  

Regards,

Robert

————————

Robert Paugh

DSPW

www.dspwi.com/contact.html

[Filipe Carvalho]

Hi.

I think that to have low latency & low BER & low power & low weight the only way is to keep away from the Wi-Fi modules and build your own dedicated wireless stack. We've done this for our wireless 9-axis motion sensor and the results are excelent. With battery included, the sensor weigths 1.9 grams.

I believe that build your own dedicated wireless stack is the way to achieve cut-edge wireless sensors.

Cheers,

Filipe

[Robert Paugh]

DSPW now has a limited number of production RCB-LVDS Wi-Fi interfaces in stock.  

Introductory price is $895.00 plus shipping.  Includes the RCB-LVDS module, a USB power cable and GUI.  Batteries, enclosure and WiFi router are not included.  Currently taking orders and shipping to the USA.  

Preliminary docs, Open API, and ap notes are available on dspwi.com.  Using the API you can control the RCB-LVDS from Matlab/Octave C#, write an open-Ephys plugin, ...

For folks in the Boston area, I offer streaming demos at the UMass Lowell iHUb  or I can visit you at your lab to demo.

Best,

Robert