A Spatial Decision Eye-Tracking Task in Patients with Prodromal and Mild Alzheimer’s Disease

Abstract

Background/Objective: Performances on spatial decision eye-tracking tasks are known to be impaired in patients with moderate Alzheimer's disease (AD), but the clinical relevance of this deficit during earlier stages of AD remains unclear. Methods: This study recruited patients with amnestic mild cognitive impairment (aMCI, prodromal AD), patients with mild AD, and age-matched controls from three French memory clinics. Participants' ability to make spatial judgments and decisions was assessed with an eye-tracking system, and cognitive performance on conventional neuropsychological tests was evaluated. Results: We enrolled 26 controls, 25 aMCI patients (median Mini-Mental State Exam [MMSE] 26), and 23 mild-AD patients (median MMSE 23). Patients with mild AD had higher error rates on the spatial decision task than aMCI patients and controls (32.4% versus 23.5%; p < 0.01 and 32.4% versus 22.2%; p < 0.05, respectively), but there were no differences among the groups in anticipation rate or the percentage of express saccades. Additionally, error rates on the spatial decision task were inversely correlated with performance on visual memory tests (immediate and delayed recall on the DMS-48: r =-0.44, p = 0.0019 and r =-0.43, p = 0.0020, respectively), semantic fluency (r =-0.44, p = 0.0016), and global cognition (MMSE: r =-0.44, p = 0.0019). Performance on the spatial decision task was not correlated with anti-saccades, processing speed, or attentional performance. Conclusions: Patients with mild AD made more errors on a spatial decision task than aMCI patients and controls. We hypothesize that impaired visuospatial judgment may explain these results and distinguish aMCI patients from mild AD patients.

Saccadic Eye Movement in Mild Cognitive Impairment and Alzheimer’s Disease: A Systematic Review and Meta-Analysis

“…Of these, eight studies (Bourgin et al, 2018 ; Bylsma et al, 1995 ; Currie et al, 1991 ; Mosimann et al, 2004 ; Pavisic et al, 2017 ; L. F. Scinto et al, 1994 ; T. Shakespeare et al, 2015 ; Verheij et al, 2012 ) did not meet the data availability inclusion criteria as the reported saccade paradigm temporal format could not be distinguished or gap and the overlap results were combined; thus, these studies were excluded from the meta-analysis. Thus, the remaining 27 studies (Abel et al, 2002 ; Alichniewicz et al, 2013 ; Boucart et al, 2014a , b ; Boxer et al, 2006 , 2012 ; Chehrehnegar et al, 2019 ; Crawford et al, 2013 , 2015 , 2019 ; de Boer et al, 2016 ; Garbutt et al, 2008 ; Hershey et al, 1983 , 2013 ; Holden et al, 2018 ; Kaufman et al, 2012 ; Laurens et al, 2019 ; Lenoble et al, 2015 , 2018 ; Mosimann et al, 2005 ; Noiret et al, 2018 ; Peltsch et al, 2014 , 2020 ; Shafiq-Antonacci et al, 2003 ; Wilcockson et al, 2019 ; Yang et al, 2011 , 2013 ) were included in the quantitative analysis (meta-analysis: see Fig. 1 ).…”

“…Of the 35 studies included in this review, 8 (23%) (Boucart et al, 2014a , b ; Bourgin et al, 2018 ; Holden et al, 2018 ; Laurens et al, 2019 ; Lenoble et al, 2015 , 2018 ; Noiret et al, 2018 ) were conducted in France, 7 (20%) (Boxer et al, 2006 , 2012 ; Bylsma, 1995 ; Garbutt et al, 2008 ; Hershey et al, 1983 ; Heuer et al, 2013 ; L. F. M. Scinto et al, 1994 ) in the United States, 8 (23%) in the United Kingdom (Crawford et al, 2013 , 2015 , 2019 ; Mosimann et al, 2005 ; Pavisic et al, 2017 ; Polden et al, 2020 ; T. Shakespeare et al, 2015 ; Wilcockson et al, 2019 ), and the rest (34%) in Australia (Abel et al, 2002 ; Currie et al, 1991 ; Shafiq-Antonacci et al, 2003 ), Germany (Alichniewicz et al, 2013 ), Canada (Kaufman et al, 2012 ; Peltsch et al, 2014 ), China (Yang et al, 2011 , 2013 ), the Netherlands (de Boer et al, 2016 ; Verheij et al, 2012 ) Switzerland (Mosimann et al, 2004 ) and Iran (Chehrehnegar et al, 2019 ). Two studies (Bylsma, 1995 ; Crawford et al, 2015 ) were longitudinal prospective cohort studies, whereas the rest were matched case–control studies.…”

“…Of the 35 studies assessed using the ROBINS-I risk of bias assessment tool (Table S4 in the Online Resource), 25 studies were rated as a moderate risk of bias (Abel et al, 2002 ; Alichniewicz et al, 2013 ; Boucart et al, 2014b ; Bourgin et al, 2018 ; Boxer et al, 2006 ; Bylsma, 1995 ; Chehrehnegar et al, 2019 ; Crawford et al, 2015 ; Crawford et al, 2019 ; Currie et al, 1991 ; de Boer et al, 2016 ; Hershey et al, 1983 ; Holden et al, 2018 ; Kaufman et al, 2012 ; Laurens et al, 2019 ; Lenoble et al, 2018 ; Mosimann et al, 2004 ; Mosimann et al, 2005 ; Peltsch et al, 2014 ; L. F. M. Scinto et al, 1994 ; Shafiq-Antonacci et al, 2003 ; T. J. Shakespeare et al, 2015 ; Verheij et al, 2012 ; Wilcockson et al, 2019 ; Yang et al, 2011 ). Ten studies were rated as having a low risk of bias (Boucart et al, 2014a ; Boxer et al, 2012 ; Crawford et al, 2013 ; Garbutt et al, 2008 ; Heuer et al, 2013 ; Lenoble et al, 2015 ; Noiret et al, 2018 ; Pavisic et al, 2017 ; Polden et al, 2020 ; Yang et al, 2013 ).…”

“…In twenty-four studies the saccade latency of patient groups (MCI and ADD) was compared with that of controls using gap conditions (Abel et al, 2002 ; Boucart et al, 2014a ; Boucart et al, 2014b ; Boxer et al, 2006 ; Boxer et al, 2012 ; Chehrehnegar et al, 2019 ; Crawford et al, 2015 ; Crawford et al, 2013 ; Crawford et al, 2019 ; de Boer et al, 2016 ; Garbutt et al, 2008 ; Heuer et al, 2013 ; Holden et al, 2018 ; Lenoble et al, 2015 ; Lenoble et al, 2018 ; Mosimann et al, 2004 ; Mosimann et al, 2005 ; Pavisic et al, 2017 ; Peltsch et al, 2014 ; Polden et al, 2020 ; T. J. Shakespeare et al, 2015 ; Wilcockson et al, 2019 ; Yang et al, 2011 ; Yang et al, 2013 ). Fourteen studies used overlap conditions (Boxer et al, 2006 ; Boxer et al, 2012 ; Chehrehnegar et al, 2019 ; Crawford et al, 2015 ; Crawford et al, 2013 ; Garbutt et al, 2008 ; Laurens et al, 2019 ; Mosimann et al, 2004 ; Mosimann et al, 2005 ; Peltsch et al, 2014 ; Polden et al, 2020 ; T. Shakespeare et al, 2015 ; Yang et al, 2011 ; Yang et al, 2013 ). Eight studies used step conditions (Abel et al, 2002 ; Alichniewicz et al, 2013 ; Currie et al, 1991 ; Hershey et al, 1983 ; Holden et al, 2018 ; Kaufman et al, 2012 ; Noiret et al, 2018 ; Shafiq-Antonacci et al, 2003 ).…”

“… Prosaccade Latency Thirty studies reported the prosaccade latency of controls compared to patients (Abel et al, 2002 ; Alichniewicz et al, 2013 ; Boucart et al, 2014a ; Boucart et al, 2014b ; Bourgin et al, 2018 ; Boxer et al, 2006 ; Boxer et al, 2012 ; Bylsma, 1995 ; Chehrehnegar et al, 2019 ; Crawford et al, 2015 ; Crawford et al, 2013 ; de Boer et al, 2016 ; Garbutt et al, 2008 ; Hershey et al, 1983 ; Heuer et al, 2013 ; Holden et al, 2018 ; Laurens et al, 2019 ; Lenoble et al, 2015 ; Lenoble et al, 2018 ; Mosimann et al, 2004 ; Mosimann et al, 2005 ; Noiret et al, 2018 ; Peltsch et al, 2014 ; Polden et al, 2020 ; L. F. M. Scinto et al, 1994 ; Shafiq-Antonacci et al, 2003 ; T. J. Shakespeare et al, 2015 ; Verheij et al, 2012 ; Yang et al, 2011 ; Yang et al, 2013 ) Nine of these studies had an MCI group alone (Alichniewicz et al, 2013 ) or with an ADD group (Chehrehnegar et al, 2019 ; Heuer et al, 2013 ; Holden et al, 2018 ; Laurens et al, 2019 ; Peltsch et al, 2014 ; Polden et al, 2020 ; Yang et al, 2011 ; Yang et al, 2013 ). Overall, 87% of the studies found that patients had a longer latency than controls, with no study reporting significantly longer latency in the control group.…”