Creating an annotated bibliography

An annotated bibliography is a list of references that also includes a brief description of each paper and the finding. In our lab students doing independent study credit are required to produce an annotated bibliography, and other students frequently compose these as well when learning a new topic. In fact, I try to make these for myself, because they are a nice way to summarize a group of articles.

There are many approaches to annotated bibliographies; I've adopted some guidelines to help with the sort of document that is most helpful in my lab:

  • Use a consistent citation format—e.g., the Journal of Neuroscience reference style.

  • Include a link to the article (for example, a doi link or link to the publisher's website).

  • Describe the article in enough detail that someone else will understand the experiment(s) and result(s), but not so much that it bogs down the document. As a rule of thumb you should be able to fit 2-3 articles on a single page of 8.5x11" (or A4) paper—these should be concise summaries.

  • Include a figure that illustrates the main finding(s) of the article. This will be useful in reminding you of the results later, but is also useful if you share your bibliography with someone else. When I review bibliographies that other people have made I find it invaluable to have a sense of what the authors' actual data and results were.

Although you can get the summaries in part from the paper's abstract, it's worth actually reading the paper and seeing what the actual results seem to show, rather than taking the authors' word for it. (Sometimes authors exaggerate!)

If you know a topic well, it is often useful to include your overall opinion of the importance or quality of the study. When you are summarizing papers that are on a new topic for you, though, it's OK to leave this out.

Examples

Below are some examples that I have made up based on the sorts of approaches people take. The example article I have summarized is our 2013 paper on brain oscillations.

EXAMPLE 1: TOO LITTLE DETAIL

Peelle JE, Gross J, Davis MH (2013) Phase-locked responses to speech in human auditory cortex are enhanced during comprehension. Cerebral Cortex 23:1378–1387.

In this paper the authors play sentences for listeners and record brain responses. They find that responses are different for intelligible speech compared to unintelligible speech.

If you know the study, this description makes sense—but someone who hasn't read the study will be confused. What kind of brain responses were involved? What did subjects have to do? How were the brain responses different?

Also, the citation looks formatted in a reasonable way, but there is no easy way for the reader to get to the paper (i.e., a link).

EXAMPLE 2: TOO MUCH DETAIL

Peelle JE, Gross J, Davis MH (2013) Phase-locked responses to speech in human auditory cortex are enhanced during comprehension. Cerebral Cortex 23:1378–1387. doi:10.1093/cercor/bhs118

In this paper the authors presented young adult listeners between the ages of 19–35 with spoken sentences. The sentences were between 2.31 and 4.52 seconds in length and were recorded by a male native speaker of British English and digitized at 22,050 Hz. For each participant, each sentence occurred once in an intelligible condition and once in an unintelligible condition. Intelligibility was manipulated using a noise vocoding technique in which the amplitude envelope of each sentence was extracted in logarithmically-spaced frequency bands, used to modulate broadband noise, and then refiltered into this same frequency band.

The authors investigated cerebro-acoustic coherence, referring to the degree to which the neural responses were phase-locked to the acoustic amplitude envelope of the speech signal. When doing this for 1 channel speech, which is unintelligible, the authors found increased coherence in many brain regions including left and right temporal cortex and premotor cortex. The significance of this effect was determined using a permutation analysis. When looking at anatomical regions of interest (ROIs) centered on bilateral auditory cortex, it is apparent that increased coherence is seen for intelligible speech (16 channels) compared to unintelligible speech (1 channel). In a whole-brain analysis of this same comparison, two clusters showed up as having more coherence for intelligible speech than unintelligible speech: one in left temporal cortex, and one in left inferior frontal cortex.

There is too much detail here to quickly digest—the detail gets in the way of getting the big picture.

EXAMPLE 3: JUST RIGHT

Peelle JE, Gross J, Davis MH (2013) Phase-locked responses to speech in human auditory cortex are enhanced during comprehension. Cerebral Cortex 23:1378–1387. doi:10.1093/cercor/bhs118

In this MEG study the authors presented listeners with spoken sentences; intelligibility was varied using noise vocoding (1 channel, 4 channels, 4 channels rotated, 16 channels). After each sentence participants repeated back as much as possible. The authors analyzed the data to see whehter phase-locked neural responses depended on intelligibilty. Phase locking was found in many regions (Figure 3 in the original paper), but increases due to intelligibility were seen in left temporal cortex and left inferior frontal cortex (Figure 4).

This is a level of description that makes it easy to tell what has been done and what the main results are and the overall task without being bogged down in too much detail.

You would get bonus points for including a picture of the results. The example below is from Figure 4 in the original paper, but I left off panel C as it makes a methodological point but is not critical to getting the main point of the paper. If someone wants to know the details, they can read the paper; the annotated bibliography is just a high-level summary.

Other sources

A quick internet search brings up many links to making an annotated bibliography, including: