 So what we're going to be doing today is supporting a structured or modular search method. I am just going to show you what we're going to do. So our plan, our methodology here, and what we will do today is use the Zora to expand and map concepts to make sure that we cover all the related words, all the synonyms, all the possible ways that these concepts exist in concept space for your discipline. And then this is something you can use definitely, you in fact have to use to expand searches to make them more sensitive, but you can also use this in precise searches. This is a very modular method. So the structured search consists of applying certain methods to make searches more sensitive by the use of synonyms from subject-specific database to Zora and connecting them using the OR operator. So once again OR means more, right? OR makes searches broader. There's some other operators you can use, but we remember that the subject-specific databases translate particularly well towards making strong sensitive searches because they have those subject-specific Zora that have terms that can truly expand their searches by introducing many disciplinary equivalents. What we're going to be using today is PubMed. So we won't be using PsyKenthal. One of the reasons for that is that PubMed does automatic mapping of terms to subject headings. So of course, PubMed uses mesh medical subject headings common to Medline, PubMed and a few other renditions of the databases have their own very specialized subject-specific specialist curated sets of terms. And we will investigate how PubMed uses the medical set to investigate our psychological research question. What we're going to be doing first is thinking about how to translate a concept map into a search and it's actually quite simple. What we do is we map each column. So here we have four concepts, four columns, gambling decisions, stimulant use, sleep deprivation and healthy adults, connect all the terms that are related to each other with ORs and then connect each of the terms once they have been fully exploded and fully explored with an ant. So the process is so simple but so important to make sure that we make reproducible, robust searches. The first thing, really, really key, search one term, one concept at a time, so important. And you can explode the concept in as many ways as you want, use the thesaurus over and over. The main thing is that finally you will come up with a way to write that concept as a bunch of ORs. You'll unite using ORs, all the words related to that concept. So you'll be mapping each of the concepts to subject headings and then including what we call text words, which are just keywords, just regular words that express the same things as the subject headings. All of this is connected with ORs. So the way this looks is as follows, you work with concept chunks. So for concept one, you go through all the subject headings from the thesaurus, you connect all the other words that are related with ORs and you leave that result alone. It could be several hundreds of thousands of results. That's fine. This is the entire concept space for your concept one. When you move on to concept two and you do exactly the same thing, you go to the thesaurus, you explore that concept fully in the thesaurus. And then once you're happy with that, you add all the variants of the same words as text words. And then you move on to concept three, do exactly the same thing. And once you've explored each of your concepts, doesn't have to be three, could be two, it could be five. What you do is you combine these concept chunks and see what is the overlap and how do you do that using ANDs? So this example here follows this question. How does the quality of gambling decisions change as a function of stimulant use for healthy sleep deprived adults? So we've got our concept map organizer and we see our four concepts that we have, the gambling decisions to stimulant use, sleep deprivation and healthy adults. Let's start mapping them. So we start with the first one with gambling decisions. And again, we're not going to be using our psych info pro quest because we want to see how PubMed will map this automatically. You'll see that the database itself uses this strategy to come up with as broad a search as it can. When we start in PubMed, we will always work with the advanced search, right? So we're going to be engaging the automatic search mapping. And what we will do is go to our databases on our homepage, find PubMed. So once again, the subject specific databases here outline in purple. And once we find PubMed, we're going to type into our advanced search window, the term gambling decisions. No quotation marks around it, just gambling decisions. One concept at a time. And what we see here on the great background is the result of the automatic term mapping that PubMed will do for us. So you see that gambling decisions got translated. You see that bolded translations term. So it got translated into everything about gambling, including the boxed gambling mesh term, and then it got translated into decisions. So gambling and decisions got split up as two terms that happened somewhere within the record. So, you know, that gives a lot of records, but a lot of them will not have anything to do with gambling and decisions as concepts related to each other. So we have to realize that. Now, how did I do this again? How did I get this? I went to the PubMed advanced search builder and that's immediately underneath the simple one on the homepage. And I just typed into the query box gambling decisions. Next thing that happened, I had this page of results. So you see it's more than 3000 results for gambling decisions, but instead of looking at them right away, instead of doing the very important appraisal part, so you remember appraisals part of every search, I actually clicked on the advanced search again, which is right underneath our display window there. So you see the red arrow points at the advanced search. And what the advanced search shows me is first the search box, but then underneath, it shows me the most important thing that subject specific databases have. Actually, a lot of these citation indexes have them as well. And that's what helps us to build searches in a structured way. The search history. So the search history that you see on the upper left-hand side is the most important thing in these advanced search boxes. The other really important thing that which is otherwise hidden is this details area. So it's right beside our search query and it has that little carrot icon. And right now you'll see the details arrow is pointing down, but that's because I toggled it. Normally it was pointing towards the right. So when you toggle the details area, it opens up and it shows us the search translation. Once I took a look at the search translations, I realized that there's no mesh thesaurus term for decision. And I thought, well, surely that would be very different if we were to use the psych info thesaurus, right? So we see both the advantages and the limitations of subject specific thesaurus. Now, of course we have to evaluate what are we going to do next? How are we going to write down our gambling term? So we see risk-taking here. Risk-taking is very broad. We get 88,000 and when we look at the results, some of them take us a little bit far off, but risk-taking is a mesh term. So we have an option to include it. I decided to exclude it and then you'll see whether or not it was a good decision. And then I went and checked a few other things like gambling behaviors, gambling addiction. And then I finally settled on writing down the gambling decisions term. So the entire concept of writing it down is gambling thesaurus term, so mesh term, or gambling decisions as a phrase or gambling behaviors. But I could have done something different, right? I could have done gambling and decisions, which means that the two terms would have happened somewhere not necessarily close together in the journal articles that we retrieved. So there are many choices. There were reasons for this choice, but you also see what this choice caused, right? In any case, we're done with gambling for now. Let's move on to our next term. And remember, we don't do anything right now. We just leave this in our search history. It's so important to remember, search histories have numbers for every search and you can always reuse that search. So you don't have to worry about anything. You just start a new search, just start writing a new search from scratch. Our next search is going to be for stimulants, has specifically stimulant use as we specified here. So let's do some automatic term mapping for stimulant use. And oh, wow, we get an exclamation point. What is going on? Well, that exclamation mark is a warning from PubMed that tells us that the word use is pretty well useless. Why? Because it happens too often in searches. So PubMed tells us it will not be mapping it because it's too common and it's just not going to help our search. So essentially we're looking for stimulant. That's fair, I like that. And look at that. Our translation shows us that stimulant has our central nervous system stimulants as a mesh term, which is great. It also has central nervous system stimulants as a pharmacological action. That's fantastic. And then it breaks up central and nervous system and stimulant into everything. And then it has a phrase central nervous stimulants. So it actually goes to town with this. It loves this term. So we're happy with that. That's good mapping. Now, of course we have a number of other terms we could use and those are amphetamines and other specific stimulants like caffeine. So let's map a few. When we map caffeine, we also find that it has its own physaurus or mesh term, which is great. And then when we map a few of the other ones, we see that all of them have mesh terms including amphetamines as a general group. So here we're including amphetamines as mesh, amphetamines plural, amphetamines single, caffeine of course, modifinal. We have all of these stimulants included. So this is a very rich search for all these stimulants and it gives us more than 150 results. And our third concept is sleep deprivation. So let's do some automatic term mapping for that. So once again, we find out that we have a mesh term for sleep deprivation, fantastic. We also have an old fields, which means a text word sleep deprivation. So we're going to use those two in this search. So we have a small number of entries for gambling. We have a smallish number for sleep deprivation and absolutely huge number for our central nervous system stimulants. So that's fantastic. Now, of course, the final step is going to be adding all three of these using ants. So we will be limiting the search using ants. Now, quick question. What do you think will happen to the number of results when we combine 6,000, 13,000 and 150,000 together with ants? So we're going to focus this search, think about the final number and let's take a look at what we're going to do. So we're going to use the Boolean operator and to combine all the chunks and to do overlap. And we got, oh wow, we only got two. And we see in our search history that the way we combined our single concept queries that we used to probe all the concepts is by using the Boolean operator and it was a search. Look how simple it is. We just use the line numbers of the searches from our search history. So it's extremely modular. It's a little bit like computer programming, but we only have two results. Wow. So this is brutal. It just shows us that most likely our gambling is very much a bottleneck and bottlenecks are common and there are strategies for dealing with bottlenecks. So let's take a look at the two that we got. They're actually very good. They are both about the effects of stimulants on decision-making in sleep deprived humans. And again, the decision-making is totally about gambling. I'm actually looking at the paper number two. It's a little bit more about decision-making than just gambling. So interesting there. So these are relevant, but how do we get more similar or related studies? There are so many ways. So first of all, look at that. Who's the first author? In both cases, it is W.D. Kilgore. So we see a lead researcher already. Key authors can often be found as the names of researchers. And again, they should be leading labs that repeat from article on the same topic from article to article, right? And we see that between 2007, that's article number two, and 2012, that's article number one, the collaborators have changed. So these are likely graduate students, post-doctoral fellows or lab assistants, and maybe even undergrad students. So you see that Kilgore stays the same. And very often the lead researcher is last, not necessarily first. So as long as you see a stable name does not have to be placed first. One of the ways of expanding on the search is to go into each of these and look at the mesh thesaurus terms that were offered for each of them. And you'll find that thesaurus terms were multiple. Many of them were much more expansive, much more general than the terms we selected. So we can see that our search selected too few terms and they were already too specific. So we see that there is even a thesaurus term decision-making forward slash drug effects. That's fantastic. We probably want that one. We are interested in neuropsychological tests. We're interested in risk-taking as a separate area. We're interested in also, well, physiological correlates or physiological study of arousal. There are so many really good mesh terms. So I would say rate these terms, include them in future iterations of your search and watch results grow. And hopefully, again, not grow out of bounds to make your searches too, too, too sensitive. So again, to keep them relatively precise. So I just want to include one more quick thing. This is the proximity operator. Proximity operators are present in some ProQuest databases. PubMed doesn't have them. So I couldn't show them to you in PubMed. But proximity operators, such as within, near, and pre, which is before, happen in a variety of databases, particularly in our citation indexes. So this is the w or within and pre or pre was sent for before, operators in Scopus. And I'll show you the syntax for using each of those. They're incredibly useful because that term in the forward slash is the number that the two terms you're searching for can be separated by. So if you do w forward slash four, whatever two terms you have, let's say caffeine and sleep deprivation, you will be able to find both of them one first, one second, doesn't matter within four terms. So there are a few rules of thumb. For example, when you do within eight, that usually covers a sentence. When you do within 12, 15, up to 28, it covers a paragraph. So there are real ways of tweaking this, especially in the citation indexes.