I'm a creature of habit, and they're generally bad. Starting with poor subject agreement, haha.

At my job, we've established an interesting routine: we have a ( closed door policy. )

Some of these are academic/professional and some are personal. 


( What I need to stop doing: )

Is very hard to balance right now.  This week was a total wash:  I find that I'm constantly up late nights trying to get work done, but it seems like what I'm studying isn't what's being asked of me.   And the fact that I'm working doesn't seem to carry any weight:  I've got to do some research, and very soon, so I'll have even less time!  

But I want to sleep and sleep and sleep.


delayed topic: modeling soluble tracers.

Because when you get to your meeting at 8, your brain is fried.  So next time...just get some sleep!

So my super-secret trick for reading scientific papers at an ungodly hour is to do the following:

1. eat a banana (sugar and potassium is good for your brain
2. drink some water (same reason as #1)
3. Read the abstracts first
4. The read the conclusions
5. Skim the methods for math you recognize

and my favorite:

6.  Use Adobe's search feature to seek out key words in the text.   That way you can better focus your reading on the papers that have the info you need, as opposed to the ocean of documents you jump in while searching Pubmed.

A good way to search Pubmed is to use the "related links" in the sidebar when you find a promising document, but another really handy trick is to use the keywords that are sometimes included by the author.  These keywords are located at the bottom of the abstract, and can make your websearch more efficient. 

I have to admit, I am a PubMed trawler:  I find myself jumping down no end of rabbit holes.

This is a slacker post, I know...tomorrow (well, later today) I'll talk more about compartment mixing.  I know it's thrilling.  :)  Really, it's just all I'm doing right now, so writing out the concepts helps me learn.  Later next week, I might post some illos, or link to my site that does.

*grousing* :  my team's project has to be reworked.  

Being anxious about your ability to your assignments can hurt your efficiency.  It might sound hokey, but the demoralizing feeling of not being able to do an assignment takes away the energy and motivation you really need to get the work done.  I tend to be super-anxious, and I find that it makes me very poky when doing assignments, because I'm second-guessing each step.

Sometimes that unease is well-founded, like when you know that you may not have the prerequisites for your class, or when you've got an overloaded schedule, or if the class is just plain kicking your behind.   In this case, a student may "clock out" and mentally gloss over the more difficult parts of the assignment.  This can lead to gaps in understanding the material as a cohesive whole.

In other cases, a person may have a feeling of perfectionism, which magnifies the importance of each assignment.  This kind of mentality may not be bad as far as your GPA, but it can be crazymaking.  And it may also lead to overworking the assignment, and missing the obvious simplifications and shortcuts that turn the material into something usable, not just a homework that's set aside after the grades are entered.

I have a tendency to do both (not good) and so I'm trying a new tactic:  reviewing my questions.  This is a quick scan of questions off of the question-recall format mentioned in the Cornell Note-Taking Method (I'll talk about it tomorrow).  The benefit to reviewing the questions is that it quickly identifies your persistent, nagging weaknesses.  So when you look at your assignments and your brain explodes, you've got a better sense of what it really is that you don't know, as opposed to other aspects of the problem (like confusing notation or odd wording) that's startling, but not actually troublesome.

The second benefit to reviewing your questions is that after you've identified what you don't know, you can be better prepared to meet with your instructor or study group.

edited:  a really great explanation is at Cal Newport's blog.  http://www.calnewport.com/blog/ 

One of the biggest challenges in an engineering or math course is being able to not only recognize, but manipulate the material presented in class.  Students may be able to identify a problem when presented exactly the same way as in the text, but not be able to understand exactly what's happening, which makes it difficult to solve anything besides the plug-and-chug style problems. 

In an engineering course, this can be devastating, because you're often asked to change parameters, make idealizations or to use math to predict the behavior of a system under different inputs.  Also, because math and science can be vertically-based learning, material from previous semesters is used, but not explained.  Right now, in a senior-level modeling class, we're using 9th grade algebra side by side with sophomore-level calculus and graduate-level biology.  And not only do we have to answer questions, we have to create our own questions, models or experiments.  So we want to use math as a tool, and we have to learn not only how to use a tool, but which one is right for the job.

An additional challenge is the workload:  a typical problem set may take 1-2 hours per problem, depending on the class.  Ours are approximately 4hrs/problem.   And because problems are given as multi-step questions, getting stuck on one part may prevent a student from finishing the rest of the question.   It's incredibly easy to burn  valuable time.

So what to do?   Unfortunately, we don't have the ability to bend time, so it's a challenge to review past material, learn new material, and combine it together to make our tool, not a noose around our throats. 

But we can make use of a few tips and tricks to help us work more effectively. 

The first trick is one I'm using:  discussion.  If you have to explain the concepts and the math to anyone, it will force you to mentally organize the information in a usable form.   And that's what's expected of engineers:  using the knowledge.  If you're discussing the actual math (the problem solving steps), it's probably more effective to talk to people in your class or near your skill level, because there's an expectation that your back knowledge is also theirs, and you won't waste time bringing someone up to speed. 

But for the concepts, it can be useful to explain them to anyone, because it forces you to be sure you really understand where the math is supposed to go (even if you're still firming up how to get there.)

So we can model the body as a set of containers, either in series (like tree lights) or in parallel (like how the fingers are oriented to the wrist.)  The easiest model makes an idealization:  all of the containers have identical volumes.  This is probably most appropriate for gas mixing, where the number of moles of gas don't vary as the volume changes.  But for an incompressible fluid, like water or blood, it's better to evaluate mixing effects with volumes similar to what's seen in different areas of the body.  This introduces a complication, because now the flow rates aren't just dependent on how much drug is injected, but on the variance of mixing based on container volume.

We can use a cooking analogy to explain this effect.  We want to make a St. Patrick's day cake, so we need to mix green dye in the batter.  We already know that given enough time, we can make sure the batter is perfectly mixed, so that all the dye we squeeze into the bowl can be accounted for if we look at the batter as we pour it in the pan.  In geek speak, we say:

concentration of dye in the container equals concentration of dye leaving the container

and we write it as [D]i =[D]o  for t@ infinity.

But what happens if we look at some time before infinity?  We're busy, and we don't have time to wait all day for the cake.  So at time =just before Lost comes on, what color is the batter we're pouring in the pan? 

It turns out that if we have a big enough container, the batter turns green, no streaks, and no dye left on the bowl:  more perfect mixing. 

But if we had split up the volume of the batter, putting it in cupcakes instead of one big cake, and keeping the pour rate constant,  then the smaller the container, the worse the mixing.  It's like a cupcake assembly line:  first we fill one cup, mix it and pour it into a second cup while trying to refill the first.  Not only does the batter not stay long enough into one cup for the stirring action to have much effect, the total volume of batter gets poured out piecemeal.  So we are left with a swirly mixture, with the dye being poorly incorporated into the batter.

In geek speak, that means we get drug dumping, instead of a nice, steady sustained release. 

The funniest thing about engineers is our need to simplify -- especially when it takes ridiculously complicated logic to do so.  I see this right now in my modeling class.  We want to look at how drugs get metabolized by the body, from the injection site all the way down to the synapses, but we don't want to think about the shape of the body, and we definitely don't want to spend hours peering through the microscope.  So we're learning how to treat the body as a set of tiny little boxes, and we want to know how do those boxes talk to each other.


So we start with the easiest kind of box:  a great big one!  And we consider that there's an input (like the mouth).  Our drug gets metabolized by the body in the compartment, and then we look at the output (blood, luckily.  My anatomy friends deal with feces...which is another vote for modeling analysis!)   

It's a straightforward model:  stuff goes in = stuff metabolized + stuff comes out.  Oddly enough, this is slightly too complicated.  We say this formally like this:

 stuff = a mass, which is found by concentration of stuff* volume  = C*V    (concentration is a mass/volume, like OJ in water)

So Cin*Vin = Cout*Vout +Creact*Vreact.   So we can see the relationships between volumes of drug and the concentrations. 

This kind of equation is called a mass balance, because we know that mass is neither created nor destroyed.  So we account for each part of it, the inputs and the outputs. 

Next post:  types of compartment models.