DELETED

(deleted)

image:

Since your frequencies are "specified", could you please share them with us?

Piezo elements are highly resonant, so they are very limited in the range of frequencies that they can emit efficiently.

You can use small speakers, which will emit sound over a wide range of frequencies, but you will need a small amplifier for each one, perhaps like this one. Speakers can't be directly connected to Arduino pins.

I found out did not measure specific frequencies

How did you try to measure frequency?

using the function tone(buzzer,250), the sound sensor would pick up a frequency of 250+-.

Using the tone function is going to inhibit any frequency measurements you make. If it were me I would use a LM567C PLL to detect the tones http://www.ti.com/product/LM567C.

But then if it were me I would know what this setup is supposed to achieve rather than what it does. In other words why do this if you know the frequency in the first place.

You do realize that there will be strong standing wave patterns in this box, right? If you're trying to do any kind of line-of-sight between transducer and microphone, you can give up right now.

I would highly suggest You get ones that can sense the undisclosed specific frequencies

Grumpy_Mike:
Using the tone function is going to inhibit any frequency measurements you make.

I'm not sure if this is what you meant, but it reminded me of another problem - tone() produces square waves, which have harmonics up the wazoo.

No I actually meant that the Tone function uses interrupts and the ISR will interfere with the constantly timed analogue readings required to detect a specific frequency.

Impulses (square wave edges ) do contain an infinite number of harmonics but those are filtered by what ever it is connected too.

I am not convinced that this project will actually achieve what the OP wants it to.

I suspect that the dome is an extremely poor model of a brain in a skull.

SMART1803:
This project is a step towards using an EEG cap. The goal is to roughly predict neuron locations in the brain. If we can accomplish that at a smaller scale, i.e. this project, then there might be some success when using the EEG cap.

@aarg Yes, I do understand, but that is the complexity of the brain.

Hopefully, this helps to clarify things for everyone.

I wonder if you could point to a reference or expand on the principles of how this is supposed to work. In particular what is it about the neuron of interest that able to be differentiated by sound waves and how is its location determined by the senor array?

It is difficult to see any resemblance whatsoever between the generation of tones and neural discharges. Neurons generate a stream of impulses, but not at any specific stable frequency so frequency analysis would be entirely useless in distinguishing one from another.

"Alpha" and such "waves" in the brain are a group phenomenon of vast collections of neurons and only their crude parameters of frequency and amplitude are estimable. Regional location in an EEG is of course traditionally done by a large array of electrodes and the impulses from discrete neurons have almost negligible propagation; it would be most unlikely that they could be detected by external electrodes at all.

(Yes, I briefly worked in neuroscience! :grinning: )

I think the reasoning is that by detecting one specific frequency using multiple sensors then you can triangulate the position of that source of frequency.
This is faulty for several reasons.

  1. First this assumes that an individual neuron produces a unique signal that can be identified by frequency, but it can’t. The average human brain has about 86 billion neurons so identifying one is simply not possible, even the site of a group of neurons can’t be identified.

but supposing it could be then
2) Simply having multiple sensors will not allow you to triangulate the source, because each sensor has no or very little direction sensitivity, you can not infer by the intensity of the signal where it has come from as even with a highly directional microphone you have a three dimensional polar diagram of sensitivity so you can’t get an angle from it. Electrodes have zero directional sensitivity. I don’t think any phase difference between each electrode is going to be enough to discriminate direction.

but suppose it could then
3) The analysis of this would ironically involve a neural network approach. That means you would have to train the system to recognise the signals and that involves knowing where the neuron group actually is.

Is this serious medical research or is it a project or assignment you are undertaking for some qualification?

My professor suggested that we would possibly use methods similar to Kalman Filtering.

Do you or your professor actually know what Kalman Filtering is or what it does. It is often banded about like it is some sort of magic bullet, it is not and I don’t see how it would help you.

This is for an independent study

So why is a Professor involved?
I used to be what is known in most of the world as a Professor, but here in the UK they are known simply as Lecturers, in fact I was a senior Lecturer, which is one up from what you probably call a professor. A professor in the UK more like a head of department grade, one up from a Reader.

This is for an independent study

So why is a Professor involved?

Well someone has to supervise, no matter how independent it is. Someone has to approve the topic and define its depth of study. Otherwise the student hands in something which might be very good, but not actually meet the criteria for his or her year of study. Good work, but sorry, that's 2nd year depth, not fourth.


Where I am it goes Junior Lecturer, Lecturer, Senior Lecturer, Associate Professor, Professor. A Professor would typically be the head of a subject, and one of the subject heads would be the HoD. So the Department of Civil Engineering would have inter alia Professors of Structural Engineering, Hydraulic Engineering, Soil Mechanics etc etc and one year the soils Prof might be the head of Civil Engineering. Full professors would certainly be at least PhDs, and often have post-doc "senior" doctorates such as a DEng where I am. Associate Profs would likely be PhDs too.

I had one Senior Lecturer who was a "mere" BSc but was nevertheless regarded as the guru in deformation of slender structures; just before retirement he finally pulled his finger out and did his Masters.

But afaik, the American model is that any lecturer gets Professor on the door?