Introduction:
We all go through the 'Why' stage as children; some of us return to it as adults. As a scientist I make a living by asking questions and trying to find their answers. For me, studying the brain and cognition is enormously compelling. It may be true that there's "nothing you can know that can't be known", but the boundary of known/unknown can be pushed. Here are some of the things I am working on and thinking about.
If you have questions or are interested in collaborating, please contact me: berryhil@psych.upenn.edu
It starts with Questions:
How do we choose what we look at, perceive and record? How do we hold on to information over the short- and long- term? How are past events indexed in the brain? What causes a memory to pop into our heads? Why do we struggle to retrieve information we need (my keys) but cannot forget things we would love to forget (say, all of 8th grade)? What brain structures serve as the neural correlates of memory? How does remembering change memories over time? What processes and neural networks allow us to shift attention seamlessly between past, present, and future elements? How can we preserve/improve memory function?
Current Research Projects:
How does Attention operate in short-term memory (STM)?
We all know that paying attention to what we are doing improves our performance, and 'fuzzing out' leads to trouble. Attention also acts on items in memory - aka being 'lost in memory'. I am investigating what permits us to shift attention to items recently placed in visual memory. For example, imagine you are trying to pick a paint to match your new, well-wrapped and out-of-view floral curtains. You may begin with the dark blue to match the irises, or decide that the dark green leaves would be better. When you get home, do the colors actually match? How were you able to compare your memory of a color with the paint chip in your hand?
This research prompts even more questions related to attention in memory: Is attention in short-term or working memory the same as perceptual attention? Is it more fragile? In one current experiment, we ask participants to remember four colored circles over a brief delay. In some trials, they are cued to one location during the delay period before they have to give a response, in other trials they are cued when a response is required. This 'retro-cue' paradigm (originally developed by Landman et al, Nobre et al) reveals that participants perform better when their attention to cued items happens before the response period regardless of whether the cue is processed automatically (e.g. rapid onset at probe location) or whether it is symbolic (e.g. '2') (Berryhill, Shay & Olson, in preparation).
What is the Mechanistic Role of the Posterior Parietal Cortex (PPC) in Memory?
Some neural structures, like the hippocampus, are canonically associated with memory. Others are relative carpet-baggers, such as the posterior parietal cortex (PPC). This area was first implicated in mnemonic processing by neuroimaging (fMRI), a powerful, but correlational method. Over the past several years in the Olson Laboratory, I have been investigating whether these activations are epiphenomenal or whether they reflect functional PPC involvement in memory. To clarify the issue, I have looked for converging evidence from a neuropsychological perspective. I have tested patients with unilateral and bilateral parietal lesions in memory tasks. Our findings indicate that PPC patients have both short-term (Berryhill & Olson, 2008ab) and long-term memory deficits (Berryhill et al., 2007; Berryhill et al., submitted a). However, these findings are nuanced and patient's performance cannot entirely be attributed to a failing in any particular stage of memory: encoding, maintenance or retrieval. For example, bilateral PPC patients are impaired when STM is probed using old/new recognition (e.g. "Did you see this picture a few moments ago") - but not when free recall probes are used (e.g. "List the pictures you saw a moment ago").In long-term memory, we found that bilateral PPC patients could not freely recall memories unless they were asked specific questions about their own autobiographical memories. We have recently found that these patients struggle to imagine future events that had not taken place (Berryhill et al., submitted b). It is thought that predicting future outcomes relies on our memory of past events. These findings and ongoing work suggests that the PPC plays an important role in assessing the quality or bringing to mind aspects of LTM and STM.
What does preserved memory function in patients indicate?
The PPC is not involved in every form of memory. It is important to identify tasks that patients perform normally so the limits of a given structure's role can be better estimated. Parietal lesions do not impair performance in implicit learning. Implicit learning happens when you do not know you are learning - here, we secretly embedded a pattern in a button-pressing task and found that PPC patients learned at the same rate as healthy adults (Berryhill, Mazuz & Olson, 2008). We next ran a study where the 'source' of information was tested. We all have source memory failings when we cannot 'place' someone, or cannot remember who told us something. In a series of source memory tasks, bilateral PPC patients performed normally when asked whether they judged a picture for pleasantness or function. However, PPC patients were less confident in their responses (Simons et al., 2009). This lack of confidence in their memories suggests they may not be able to fully re-experience past events as well as controls. This possibility opens new questions and is an area we will be investigating in the future.
How do the neural structures responsible differ for short- and long- term forms of memory?:
There is a deep canyon dividing short- and long- term memory. The rationale for this division can now be examined with a variety of new experimental approaches. Using MRI, I have been investigating cortical regions commonly activated during STM and LTM tasks. Although commonly activated, different networks participate with different levels of activity depending on the task. In one current study, participants encoded word lists before scanning. During their scan, subjects viewed words and assessed whether they were previously encoded or new (LTM task). Subjects also performed an STM task in which four colored triangles flashed. After a brief delay, a single colored triangle reappeared and subjects determined whether this probe triangle had changed color. These tasks allow us to examine shared and unique patterns of activity for each type of memory task.
When someone walks towards you, you do not have the percept that the person is growing with every step. This process of estimating size and distance (size constancy) is so automatic and seemingly trivial, that we fail to appreciate the intricacy of the process. However, the visual system, specifically areas along the occipital-parietal lobe known as the dorsal stream, integrate a wide array of information from diverse sources to determine distance. Indeed, there is a long list of monocular and binocular distance cues including stereopsis, linear perspective, occlusion, shadow, motion parallax, atmospheric haze, familiar size, etc. I have been really interested in determining how we develop this 3-D representation from pictures - so that distance information is only implied. In several studies, I have been using adaptation-MRI paradigms to understand what structures are responsible for tracking changes in distance (Berryhill & Olson, in revision). In one study with one of our amazing PPC lesion patients, we found that superior occipital areas were needed for interpreting distance in 3-D (real objects) or 2-D (pictures) (Berryhill, Fendrich, & Olson, 2009). The manner in which distance cues are integrated, how conflicting information is mediated, the neural architecture of distance information are all open questions.
The figure shows anatomical MRI brain scans of two extremely rare patients with bilateral PPC damage. The lighter spots (hypodensities) indicate areas of damage. For both patients, their primary symptom is simultanagnosia: the inability to attend to more than one object at a time. This symptom is most often associated with Balint's syndrome.
Future Directions:
Every experiment opens new questions - many of them intriguing and of theoretical importance. While reserving the right to use my brain to change my mind, I want to pursue a primary line of research investigating mnemonic attention. As an example of where I want to go, first remind yourself that you are a continual time-traveler: revisiting past events, mental states, thought processes and feelings. We use memory flexibly to guide future outcomes. I am interested in studying the interplay between perceptual attention in the present (our current physical reality) and mnemonic attention (attention to relevant aspects of past events). The very same memory of reading an article may be used to remember the great idea you meant to pursue, or the response your boss had when you mentioned it to her, or perhaps more importantly - where you were the last time you had your glasses.