Elsevier

Neurobiology of Aging

Volume 34, Issue 12, December 2013, Pages 2759-2767
Neurobiology of Aging

Regular article
Age-related dedifferentiation and compensatory changes in the functional network underlying face processing

https://doi.org/10.1016/j.neurobiolaging.2013.06.016Get rights and content

Abstract

Recent evidence has shown that older adults fail to show adaptation in the right fusiform gyrus (FG) to the same face presented repeatedly, despite accurate detection of the previously presented face. We used functional magnetic resonance imaging to investigate whether this phenomenon is associated with age-related reductions in face specificity in brain activity and whether older adults compensate for these face-processing deficiencies by increasing activity in other areas within the face-processing network, or outside this network. A comparison of brain activity across multiple stimulus categories showed that, unlike young adults who engaged a number of brain regions specific to face processing, older adults generalized these patterns of activity to objects and houses. Also, young adults showed functional connectivity between the right FG and its homologous region during face processing, whereas older adults did not engage the left FG but showed a functional connection between the right FG and left orbitofrontal cortex. Finally, this frontotemporal functional connection was activated more strongly in older adults who performed better on a face-matching task (done outside of the scanner), suggesting increased involvement of this functional link for successful face recognition with increasing age. These findings suggest that 2 neural mechanisms, dedifferentiation and compensatory neural recruitment, underlie age differences in face processing.

Introduction

Recent neuroimaging studies have yielded evidence for 2 distinct phenomena in the aging brain: (1) neural representations of different cognitive processes become less selective and their neural signature less distinct (Li et al., 2001); and (2) older adults often have greater brain activity than young adults during cognitive tasks, particularly in the frontal cortex, that might compensate for age-related processing deficiencies (Cabeza et al., 2002, Duverne et al., 2009, Grady, 2012, Reuter-Lorenz et al., 2000). The first finding is consistent with the idea of dedifferentiation, which in terms of brain activity refers to reduced distinctiveness of neural representations in domain-specific areas. For instance, ventral occipital visual areas show reduced category selectivity to faces, places, and words in older relative to younger adults (Park et al., 2004). Dedifferentiation has also been reported in domain-general cognitive areas, including the parietal and prefrontal cortices, during memory encoding and retrieval (Carp et al., 2010a, Carp et al., 2010b), and in areas active specifically for retrieval of autobiographical and episodic memories (St-Laurent et al., 2011). Although such results might indicate reductions in the integrity of the aging brain, other neuroimaging studies have reported an increased engagement of prefrontal and other brain areas, which is interpreted as a compensatory mechanism when associated with maintained performance in older adults (Davis et al., 2008, Grady, 2002, Grady et al., 2002, Grady et al., 1994, Madden et al., 2004, Schiavetto et al., 2002), or when activity in these “over-recruited” areas is correlated with behavior in older adults (Davis et al., 2008; Grady et al., 2005). Indeed, some have suggested that the strongest evidence for compensation is this latter finding, in which a link can be made between more brain activity and better performance in older adults who show the most overrecruitment (Cabeza and Dennis, 2012; Grady, 2008). Albeit 2 distinct phenomena, the dedifferentiation and compensation processes are unlikely to be mutually exclusive. Rather, it is possible that the brain might show reduced neural selectivity in some domain-specific regions and, at the same time or as a consequence, use other task-specific regions, or even a different network of areas, to compensate for this deficiency in neural distinctiveness, evincing remarkable brain plasticity or reserve in old age. The purpose of the current study was to investigate this question by exploring age-related neural changes during face processing, using functional magnetic resonance imaging (fMRI).

Many behavioral studies have reported that older adults show reduced recognition and perception of familiar and unfamiliar faces relative to younger adults (Bartlett et al., 1991, Boutet and Faubert, 2006, Habak et al., 2008, Lott et al., 2005, Searcy et al., 1999). On the neural level, age-related deficits in face processing are likely to be observed in multiple areas (Carp et al., 2010a, Goh et al., 2010, Lee et al., 2011), because face processing is distributed widely over many functionally interacting areas that show serial and parallel processing (Barbeau et al., 2008, Gobbini and Haxby, 2007, Haxby et al., 2000). Importantly, deficiencies at early stages of face processing might cascade downstream and yield changes in the entire functional brain network (Davis et al., 2008, Grady, 2000, Grady, 2008). Thus, we aimed to examine (1) whole-brain activations during processing of faces and objects, to assess neural selectivity to different categories of objects; and (2) functional connectivity of the distributed face processing network, using multivariate partial least squares (PLS) analysis.

In our recent study (Lee et al., 2011), older adults showed no adaptation in the right fusiform gyrus (FG), 1 of the regions considered to be crucial for face recognition (Clarke et al., 1997, Kanwisher et al., 1997, Nestor et al., 2011, Steeves et al., 2006), to repeatedly presented faces, even when facial identity and view were kept constant. Contrary to these deficiencies in neural adaptation, on a behavioral level older adults performed similarly to young adults in matching the same facial identity shown in the same view outside the scanner (also consistent with previous behavioral data of Habak et al., 2008 and Searcy et al., 1999). Additionally, we found that older adults recruited a unique set of brain regions in which activity correlated with their behavioral performance. It has been suggested that older adults compensate for processing deficits because of decreased activity in the occipital lobe by increasing frontal activity (Davis et al., 2008, Grady et al., 2002). If compensation occurs, 1 possible frontal area of compensatory recruitment might be the orbitofrontal cortex (OFC), in which activity has been often observed during processing of faces and nonfacial objects (Bar, 2009, Bar et al., 2006, Fairhall and Ishai, 2007, Ishai, 2007, Ishai, 2008, Johnson, 2005, Kveraga et al., 2007, Li et al., 2010). The OFC is argued to be a part of the extended cortical network for face processing (Fairhall and Ishai, 2007, Haxby et al., 2000) and is involved in a variety of face tasks, including assessment of facial attractiveness (Ishai, 2007), facial sex categorization (Freeman et al., 2010), facial emotion recognition (Harmer et al., 2001), and detection of blurred faces (Li et al., 2010, Summerfield et al., 2006). More activity or stronger functional connectivity in the OFC in older adults would be consistent with the compensation-related utilization of neural circuits hypothesis (Reuter-Lorenz and Cappell, 2008). The compensation-related utilization of neural circuits hypothesis is based on the idea that as task demands increase, reliance on neural resources increases regardless of age, but that this demand/resource function is shifted to the left in older adults. That is, older adults recruit more resources at lower levels of cognitive load. For basic face recognition, this increased reliance on neural resources could involve task-specific regions, such as the OFC, which younger adults might only recruit when the cognitive demands are greater or more complex processing of faces is required. Additional domain-general resources might also be recruited by older adults. Both of these types of recruitment could be compensatory, especially if this additional engagement of brain activity was associated with improved performance in older adults (Grady, 2012).

To measure functional connectivity between the frontal cortex and the fusiform areas, we identified seed regions in the fusiform gyri and in the left medial OFC. Activity in this latter area has been found during viewing of famous and emotional faces in young adults (Fairhall and Ishai, 2007, Ishai et al., 2005). We expected that age-related deficiencies at early processing stages (i.e., in the FG) would cascade downstream and alter the face-processing network (Davis et al., 2008, Grady, 2002, Grady, 2008). We hypothesized that: (1) older adults would exhibit a general decrease in neural specificity across activated face-processing areas; (2) young and older adults would show differences in the functional connectivity of the face-processing network, with older adults showing stronger connectivity with OFC; and (3) activity in the face-processing network specific to older adults would correlate with their behavioral performance (i.e., reaction times [RTs] and accuracy), reflecting the importance of the OFC in face-matching performance with increasing age.

Section snippets

Participants

Fifteen healthy older adults (mean age, 68 years; SD, 4.2; range, 61–75 years; 6 men) and 14 healthy young adults (mean age, 24 years; SD, 4.9; range, 8–32 years; 7 men) participated in the study. Older adults were screened via a detailed phone interview for general health (e.g., cardiovascular disease or stroke), medications, and normal vision (e.g., cataract or glaucoma, eye exam within a year). They achieved an average score of 29.3 (SD, 1.1) on the Mini-Mental State Examination (Folstein

Behavioral performance

In both behavioral tasks, there was no significant difference in either accuracy or RT between young and older adults. In the scanner (same/different task), older adults' RT was 814.87 ms (standard error [SE] = 38.09) and accuracy 0.95 (SE = 0.02), whereas young adults' RT was 739.57 ms (SE = 28.92) and accuracy 0.95 (SE = 0.01). Outside the scanner (facial identity-matching task), older adults' RT was 797.50 ms (SE 47.94) and accuracy 0.97 (SE = 0.01), whereas young adults' RT was 708.87 ms

Discussion

In our recent study (Lee et al., 2011), we found that when viewing repeatedly presented faces of same identity and viewpoint, in contrast to their young counterparts, older adults failed to show adaptation in the right FG, despite accurate face matching. The purpose of the current study was to examine further how age-related differences in the neural underpinnings of face processing influence behavior. Our findings provide evidence for the involvement of 2 critical phenomena, a generalized

Disclosure statement

All authors have no conflicts of interest.

The study and all experimental procedures were conducted with the approval of the Baycrest Centre Research Ethics Board.

Acknowledgements

This study was supported by a Canadian Institutes of Health Research grant to M.M. and C.L.G. (MOP106301). C.L.G. is also supported by the Canada Research Chairs program, the Ontario Research Fund, and the Canadian Foundation for Innovation.

References (70)

  • J.V. Haxby et al.

    The distributed human neural system for face perception

    Trends Cogn. Sci.

    (2000)
  • A. Ishai

    Sex, beauty and the orbitofrontal cortex

    Int. J. Psychophysiol.

    (2007)
  • A. Ishai

    Let's face it: it's a cortical network

    Neuroimage

    (2008)
  • A. Ishai et al.

    Face perception is mediated by a distributed cortical network

    Brain Res. Bull.

    (2005)
  • A. Krishnan et al.

    Partial least squares (PLS) methods for neuroimaging: a tutorial and review

    Neuroimage

    (2011)
  • J. Li et al.

    Effective connectivities of cortical regions for top-down face processing: a dynamic causal modeling study

    Brain Res.

    (2010)
  • S.C. Li et al.

    Aging cognition – from neuromodulation to representation

    Trends Cogn. Sci.

    (2001)
  • A.R. McIntosh et al.

    Spatial pattern analysis of functional brain images using partial least squares

    Neuroimage

    (1996)
  • A.R. McIntosh et al.

    Spatiotemporal analysis of event-related fMRI data using partial least squares

    Neuroimage

    (2004)
  • P.D. Sampson et al.

    Neurobehavioral effects of prenatal alcohol: part II. Partial least squares analysis

    Neurotoxicol. Teratol.

    (1989)
  • A. Schiavetto et al.

    Neural correlates of memory for object identity and object location: effects of aging

    Neuropsychologia

    (2002)
  • J.K. Steeves et al.

    The fusiform face area is not sufficient for face recognition: evidence from a patient with dense prosopagnosia and no occipital face area

    Neuropsychologia

    (2006)
  • H.R. Wilson et al.

    Visual bandwidths for face orientation increase during healthy aging

    Vision Res.

    (2011)
  • M. Bar

    The proactive brain: memory for predictions

    Philos. Trans. Roy. Soc. B Biol. Sci.

    (2009)
  • M. Bar et al.

    Top-down facilitation of visual recognition

    Proc. Natl. Acad. Sci. U.S.A

    (2006)
  • E.J. Barbeau et al.

    Spatio temporal dynamics of face recognition

    Cereb. Cortex

    (2008)
  • J.C. Bartlett et al.

    False recency and false fame of faces in young adulthood and old age

    Mem. Cogn.

    (1991)
  • I. Boutet et al.

    Recognition of faces and complex objects in younger and older adults

    Mem. Cogn.

    (2006)
  • R. Cabeza et al.

    Frontal lobes and aging: deterioration and compensation

  • J. Carp et al.

    Age differences in the neural representation of working memory revealed by multi-voxel pattern analysis

    Front. Hum. Neurosci.

    (2010)
  • J. Carp et al.

    Age differences in neural distinctiveness revealed by multi-voxel pattern analysis

    Neuroimage

    (2010)
  • S.W. Davis et al.

    Que PASA? The posterior-anterior shift in aging

    Cereb. Cortex

    (2008)
  • V. Della-Maggiore et al.

    Corticolimbic interactions associated with performance on a short-term memory task are modified by age

    J. Neurosci.

    (2000)
  • S. Duverne et al.

    The relationship between aging, performance, and the neural correlates of successful memory encoding

    Cereb. Cortex

    (2009)
  • B. Efron et al.

    The bootstrap method for assessing statistical accuracy

    Behaviormetrika

    (1985)
  • Cited by (50)

    • Resting-state functional brain connectivity predicts cognitive performance: An exploratory study on a time-based prospective memory task

      2021, Behavioural Brain Research
      Citation Excerpt :

      Yet more, we interpret the age-related reduction in modularity in the OE condition (i.e. with individuals at rest but engaged in sensory/perceptual processes) as supporting the dedifferentiation hypothesis in aging [28,62]. Older adults would indeed activate unspecific neural regions, losing the distinctiveness of younger adults’ neural representations in cerebral areas that code for specific information [63], which are also involved in OE resting-state networks. Accordingly, in the OE condition we found an increased fronto-parietal midline connectivity in older participants, which is consistent with previous research [49] and suggests a compensatory effect for a decline in bottom-up sensory input processing [49,64,65].

    • Disentangling the effects of age and mild traumatic brain injury on brain network connectivity: A resting state fMRI study

      2021, NeuroImage: Clinical
      Citation Excerpt :

      Particularly, increased connectivity involving frontal areas (e.g. PFC), which are associated with high cognitive processing, seems to reflect an age-related compensation phenomenon that was termed ‘Posterior to Anterior Shift in Ageing’ (PASA; Davis et al., 2008). Dedifferentiation and compensation are unlikely to be mutually exclusive (Burianová et al., 2013; Geerligs et al., 2015). Because the current study used resting-state fMRI, in contrast to task-based fMRI, it is not possible to provide direct evidence of which mechanism (dedifferentiation and/or compensation) plays a stronger role in our sample population regarding changes in connectivity.

    View all citing articles on Scopus
    View full text