Brain Scans Analyzed Differently By PC and Macs

H

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
If you had the results of your last MRI or CT scan analyzed by a PC, your have nothing to worry about. If the same scan was analyzed by a Mac, we've got bad news for you. :D

German researchers Ed Gronenschild and colleagues took a set of 30 brains and got FreeSurfer to estimate the size and thickness of various structures. Then they did the same thing, on the exact same brains, with a different version of the software. They found substantial differences in regional volumes, depending upon the version of FreeSurfer used. Running the same version of the software on a Mac vs a PC also created differences, and even the version of Mac OS had an impact.
Click to expand...
 
It's only Monday and you're already kicking over the Apple vs PC ant hill, Steve.
 
It's a software thing, not a Mac vs PC thing. Nothing to see here, move along...
 
I only skimmed through the article, the image included shows one was running version "4.3.1" and the other was running version "5.0.0". Without know the versioning between the MAC version vs the PC version, i'm not at all surprised. theyre dealing with two different major versions. perhaps the algorithms used in one version are more accurate than the other.
 
Cliffs Notes: The latest and likely most widely distributed/tested version of the software is for PC. The Mac version, like most "afterthought" ports, has issues.

Lesson: Keep your software up to date and if everyone is using a PC, you should too.
 
Talking about making something out of nothing. Buggy code is buggy. Turning it into a PC vs MAC thing is just infantile.
 
Monkey God said:
Talking about making something out of nothing. Buggy code is buggy. Turning it into a PC vs MAC thing is just infantile.
Click to expand...

Running the same version of the software on a Mac vs a PC also created differences, and even the version of Mac OS had an impact.

Same version of software and difference version of Mac OS changed results.

Same version of software used on pc and mac, changed the results.

Gonna go with it is a pc vs mac issue along with bad programming. You should expect the same version of the program to work the same regardless of the OS version, which it didn't.
 
If they wanted to be thorough they should have tested different versions on different Windows versions as well.
 
Dekar12 said:
Gonna go with it is a pc vs mac issue along with bad programming. You should expect the same version of the program to work the same regardless of the OS version, which it didn't.
Click to expand...

No *I* don't, because I know how programming on different operating systems works. You don't just check out the code and switch a compiler setting for another OS and compile it.
 
Dekar12 said:
Gonna go with it is a pc vs mac issue along with bad programming. You should expect the same version of the program to work the same regardless of the OS version, which it didn't.
Click to expand...
It seems pretty clear that the software maker is relying on APIs provided by each respective OS (at least on the Mac side), and the implementations behind those APIs are being modified between OS versions. You could, perhaps, argue that this is a bad thing, but if the software maker isn't doing unit tests or assertions or any kind of result validation-based testing on their software, that's what's going to happen.

There's nothing stopping these guys from producing software that yields identical results regardless of platform. Given the nature of the software, I can't even imagine why this wouldn't be a goalpost for these guys.
 
ml_paladin888 said:
I only skimmed through the article, the image included shows one was running version "4.3.1" and the other was running version "5.0.0". Without know the versioning between the MAC version vs the PC version, i'm not at all surprised. theyre dealing with two different major versions. perhaps the algorithms used in one version are more accurate than the other.
Click to expand...

There were comparisons of 4.3.1, 4.5.0, and 5.0.0 on the respective platforms from what I gathered from the journal article. The biggest differences were of course on cross version comparisons. Differences between platforms on equivalent software versions were present on the two older versions but no significant differences were present on 5.0.0.
 
When's the next "I'm a Mac and I'm a PC" ad going to come out and will this study be used? :p
 
Monkey God said:
Oh by the way, the "PC" they used was running CentOS 5.3.
Click to expand...

IMO, the acronym PC really needs to be dropped anyway. There was no historical link between the term PC and the Windows OS, so there is no reason for it to specifically mean Windows today, and given that Apple hardware is now basically the same and can run Windows or Linux too, it would apply to that as well.

I don't really feel strongly about it or anything but your post definitely pointed out how the term can be used oddly.
 
Apple people get really upset when I say stuff like, "Oh my gawd, you just got a new Apple PC!" or "I live the new Macbook PCs!" I do it on purpose.
 
To call this mac vs PC is bogus, this is simply shit software that was unvalidated and should never be used for medical purposes. I used to write FDA regulated software for medical devices that analyzed data and I can tell you that excuse would never fly during an audit. It's up to the vendor to check that critical outputs work exactly the same across versions and platforms, and if they don't, they better have intended for there to be a difference and the doctors better know what it is and why.

Cliff's Notes: lazy developers ignore regulations.
 
Coming from someone who has had brain surgery - I can already say that this would make the biggest difference ever to me.
 
MavericK96 said:
From my experience very few businesses actually use Macs.
Click to expand...
Possibly, on the low end where IT doesn't have to pay attention to what employees want this is true. IT doesn't like Macs because they're different and hard to manage. Also, the cheap end of IT guys are almost all mass produced MCSE types who don't know what NFS, LDAP or SSH stand for.

It is my experience that high value employees that are competent to handle their own IT such as software, electrical, aerospace engineers often use Macbook Pros. The hardware is sweet and work will often pay for it so why not?

Where I work, I have a 2008 Macbook Pro as a secondary machine. I replaced it with a Samsung last year. Another software guy has a new Macbook Pro, and another tech guy has a 2010 model. That is 3 Macs out of 7 tech guys. 3 out of 12 if you count the management and financial types.

So anyway: In my experience Macbook Pros show up quite often in businesses. Not any other kind of Mac, oddly. Just the laptops, iPads and iPhones.
 
That is 100% programmer FAIL! The company should fire their CIO I all comes from the top.
 
I wish I could edit here.

Something [H] needs to do, upgrade the forums.
 
Interesting to see this. I recently began working as an undergraduate student volunteer in a Neuroscience laboratory. One of the studies we're performing on Aphasic patients deals with fMRI images, which we process in Linux using the AFNI software suite. Before we can process the images, they must be renamed and reordered in Matlab, but the we only have an older windows version.

The previous research assistant who set up the protocol for processing the fMRI data left, and not really anyone there understands the process. I've spent the last month setting up the systems according to their protocol (with no experience), and recently began to wonder if, between Windows and Linux, there might be some sort of discrepancy. I recently secured us a MATLAB license for 2012a (Windows AND Linux) so I guess I'll see. :)

THANKFULLY we won't run into any of these issues because we won't be using SPM8--the HDR/IMG (3D Analyze( format isn't what we need.

Anyone who is involved in this line of work should consider using Oxford's FMRIB software suite. It looks promising.
 
ol1bit said:
I wish I could edit here.

Something [H] needs to do, upgrade the forums.
Click to expand...

It's not a technical limitation, it's policy. Since front page news has the potential to be contentious topics, it's no-edit to make people think about what they say before they say it.
 
Zan Lynx said:
It is my experience that high value employees that are competent to handle their own IT such as software, electrical, aerospace engineers often use Macbook Pros. The hardware is sweet and work will often pay for it so why not?
Click to expand...

Ethics. The ethics of buying something for its name-brand value/cool factor when in several of the cases described is paid for by tax payers and they might feel being public servants, they have a responsibile to the public they serve. You know, how professionals are suppose to have ethics. Hence why some people in this thread have already mentioned, if they say their doctor using a Mac, they'd be tempted to switch offices. Nothing says, 'I am a logical, concise thinker' quite like 'I spent three times the amount of hardware as I would versus PC to get a slightly slimmer notebook and look forward to less software choices in the future but my 13 year old teenager neighbour idealizes me'.

Honestly, my experience has been the exact opposite with 'electrical enginering and aerospace design' specialists I've ever meet which would admittedly be a small percentage. The specialized software they needed was made for Windows and they ran Windows. Yes, they know about bootcamp and other work-arounds but didn't want a work-around to do their 'work'. Get it?
 
There are a great many things that Mac OS is not suited to do. CAD, CNC, Medical, and most serious business desktop use due to the relatively poor handling of domain controllers, active directory etc. It's not the swiss army knife that is windows because the business applications developed for it are often not as mature.

I think that's OK. It doesn't have to be those things. People like seem to like it the way it is. I'm just not one of them.
 
ol1bit said:
That is 100% programmer FAIL! The company should fire their CIO I all comes from the top.
Click to expand...

Indeed. I think this comes down to using native common libraries. If you want something to run cross-platform you write your own library and test the results to make sure nothing goes wrong, otherwise you can have floating point truncation, rounding, and other behaviors you do not want in statistical calculations. This is critical if you're feeding results back through the same functions that generated them as you'll have compounding precision errors. Write one library, stick with it, use as little native code as possible, and watch your floating point precision. The only thing you want native is I/O, memory allocation, and clock functions. All your math should be sandboxed.
 
MavericK96 said:
From my experience very few businesses actually use Macs.
Click to expand...

Every local doctor I've been to, including my dentist, my ophthalmologist, and the local orthopedic center, all use Windows XP on their office systems. Doesn't that make you feel warm and fuzzy?

Of course, the reason for this absurd state of affairs is that they all use vertical market office management and medical software that requires XP. Why the vendors don't update their wares remains a mystery, but I'd feel much better about my patient data were the office using Win 7, Macs, or even Linux boxes.
 
WBurchnall said:
Ethics. The ethics of buying something for its name-brand value/cool factor when in several of the cases described is paid for by tax payers and they might feel being public servants, they have a responsibile to the public they serve. You know, how professionals are suppose to have ethics. Hence why some people in this thread have already mentioned, if they say their doctor using a Mac, they'd be tempted to switch offices. Nothing says, 'I am a logical, concise thinker' quite like 'I spent three times the amount of hardware as I would versus PC to get a slightly slimmer notebook and look forward to less software choices in the future but my 13 year old teenager neighbour idealizes me'.
Click to expand...

Are you seriously suggesting that a doctor's choosing to use a Mac is unethical? I gotta say, that's a new one on me.

I find it amazing, now that Apple's the largest tech company in the world, that some people still hold the view that only idiots buy and use Macs. Whatever helps you feel better about your bloatware-stuffed commodity PC, I guess. At least it was cheap.
 
You must log in or register to reply here.
Back
Top